Intelligent emotion pacifying system, method and equipment in noise environment

By designing an intelligent emotional soothing system in a noise environment, using task matching, sound recognition and emotion acquisition modules, combining neural networks and hidden Markov models, identifying and processing noise, providing emotional soothing, the operator solves the long-term long-term operation of a noise environment. The stress and mental health problems caused by exposure have achieved effective emotional comfort and improvement of work experience.

CN119049515APending Publication Date: 2024-11-29ZHEJIANG YUNFENG MOGANSHAN FURNITURE CO LTD +1
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411092722.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-09
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

In noisy environments, prolonged exposure of operators leads to hearing fatigue, increased stress and mental health problems, and prior art has failed to effectively address these problems.

Method used

An intelligent emotional soothing system was designed, including a task and emotion matching module, a sound input module, an emotion acquisition module, an emotion soothing module and a sound output module. Through task complexity matching, sound classification, emotion recognition and noise processing, emotional soothing is provided. , uses neural networks and hidden Markov models to identify sound features, combines skin electrosensitivity to identify emotions, and outputs positive emotion-oriented through headphones and bracelet devices.

Benefits of technology

It effectively improves the operator's emotional comfort, reduces stress, and improves work experience in a noisy environment. It is suitable for a variety of working environments and is highly practical and can deal with the emotional needs of different operators.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119049515A_ABST
    Figure CN119049515A_ABST
Patent Text Reader

Abstract

The invention discloses an intelligent emotion pacifying system, method and device in a noise environment, and relates to the technical field of emotion recognition, the intelligent emotion pacifying system in the noise environment comprises a task and emotion matching module, a sound input module, an emotion acquisition module, an emotion pacifying module and a sound output module; the task and emotion matching module is used for generating a high-emotion task and a low-emotion task; the sound input module is used for collecting sound information in the high-emotion task and performing classification processing on the sound information; the emotion collection module is used for collecting emotion types and emotion change trends in the low-emotion tasks and carrying out emotion recognition; and the emotion pacifying module is used for sequentially performing noise reduction and sound conversion processing on the environmental noise in combination with the emotion recognition result and the sound information classification result. According to the invention, emotion pacifying processing can be effectively carried out according to emotion demands of different operators, the working experience is improved, and negative emotion pacifying is carried out.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of emotion recognition, and in particular to an intelligent emotion soothing system, method and device in a noisy environment. Background Art

[0002] In factory environments and other noisy environments, operators often face work discomfort caused by long-term exposure to noise, which may include hearing fatigue, increased stress, and frustration due to difficulty in communication. These factors not only affect employees' work efficiency, but may also affect their long-term health and psychological state.

[0003] In factory environments and other noisy environments, long-term noise exposure has a serious impact on the physiological and mental health of operators. Therefore, the importance of intelligent emotional soothing systems is particularly prominent. By monitoring the emotional state of operators in real time and providing timely emotional intervention, such as playing soothing music, providing relaxation tips or emotional support messages, helping employees manage stress and anxiety, this instant feedback and support is essential to maintaining the mental health of operators, especially in stressful working environments.

[0004] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0005] In view of the problems in the related art, the present invention proposes an intelligent emotion soothing system, method and device in a noisy environment to overcome the above-mentioned technical problems existing in the existing related art.

[0006] To this end, the specific technical solution adopted by the present invention is as follows:

[0007] According to a first aspect of the present invention, there is provided an intelligent emotion soothing system in a noisy environment, the intelligent emotion soothing system in a noisy environment comprising a task and emotion matching module, a sound input module, an emotion collection module, an emotion soothing module and a sound output module;

[0008] The task and emotion matching module is used to match tasks with high emotions and low emotions based on task complexity, generating high emotion tasks and low emotion tasks;

[0009] The sound input module is used to collect sound information in high-emotion tasks and classify and process the sound information;

[0010] Emotion collection module, used to collect emotion types and emotion change trends in low-emotion tasks and perform emotion recognition;

[0011] The emotion soothing module is used to combine the emotion recognition results and the sound information classification results to perform noise reduction and sound conversion processing on the ambient noise in sequence, so as to achieve the soothing effect on negative emotions;

[0012] The sound output module is used to output sound signals to human ears and provide positive emotional guidance through emotional soothing equipment.

[0013] Preferably, the task and emotion matching module matches the task with high emotion and low emotion based on the task complexity to generate the high emotion task and the low emotion task, including:

[0014] Calculate the task complexity and compare the task complexity with a preset threshold;

[0015] The calculation formula of task complexity is:

[0016] Task Complexity Score(TCS)=∑(w i × i );

[0017] In the formula, w i represents the weight of the i-th factor; r i represents the quantitative score of the i-th factor; TCS represents task complexity;

[0018] According to the comparison between task complexity and preset threshold, the task attributes are divided into high emotion tasks, medium emotion tasks and low emotion tasks;

[0019] The cognitive load of the medium-emotion task is measured, and the medium-emotion task is divided into a high-emotion task and a low-emotion task according to the comparison result of the cognitive load measurement result with the preset threshold;

[0020] High emotion tasks are sent to the sound input module, and low emotion tasks are sent to the emotion collection module.

[0021] Preferably, the sound input module includes an environmental sound collection module, a sound screening and filtering module, a sound recognition module and a sound classification module;

[0022] Environmental sound collection module, used to collect sound information in high-emotion tasks using sound receiving equipment;

[0023] A sound screening and filtering module is used to extract a set of sound types that meet a preset range from the sound information and filter out the sound information outside the preset range;

[0024] A sound recognition module, used to sequentially recognize the sound types and sound elements of the extracted sound type set;

[0025] The sound classification module is used to distinguish the sound features in the sound information using a neural network model.

[0026] Preferably, the sound classification module comprises, when distinguishing the sound features in the sound information using the neural network model:

[0027] Use Fourier transform technology to obtain sound frequency components from sound information;

[0028] Calculate the average energy of the sound signal based on the sound frequency components to obtain the sound intensity; use frequency analysis technology to determine the sound resonance frequency to obtain the timbre characteristics; use Mel frequency cepstrum coefficients to extract speech features;

[0029] The sound intensity is converted into an observation symbol set, and the influence of noise in the sound intensity is identified by using a hidden Markov model, and the noise influence is removed from the sound intensity;

[0030] The sound intensity, timbre characteristics and speech characteristics after eliminating the influence of noise are used as the input of the neural network model respectively, and the neural network model is used to distinguish the human voice characteristics and machine voice characteristics in the sound information.

[0031] Preferably, converting the sound intensity into an observation symbol set and using a hidden Markov model to identify the influence of clutter in the sound intensity comprises the following steps:

[0032] Mapping the sound intensity into a finite set of observation symbols, wherein each observation symbol in the set of observation symbols represents a corresponding sound intensity level;

[0033] The observation symbols in the observation symbol set are combined into a propagation trajectory, and the propagation trajectory is used as an observation value sequence of the hidden Markov model;

[0034] The initial state transfer probability and initial observation probability are set for the hidden Markov model, and the parameters of the hidden Markov model are gradually adjusted using an iterative optimization algorithm to obtain an optimized hidden Markov model;

[0035] The optimized hidden Markov model is used to identify the clutter effect in sound intensity.

[0036] Preferably, the hidden Markov model parameters are gradually adjusted using an iterative optimization algorithm to obtain an optimized hidden Markov model, which includes the following steps:

[0037] Extracting the initial parameters of the hidden Markov model and calculating the forward probability and backward probability of the hidden Markov model at a preset time point, wherein the initial parameters are the initial state transition probability and the initial observation probability;

[0038] Combine the forward and backward probabilities to calculate the occupancy probability of a single state at any time point in the observation sequence and the joint probability of two consecutive states;

[0039] Calculate the expected value of the joint probability of the two consecutive states based on the joint probability of the two consecutive states, and update the initial parameters of the hidden Markov model based on the expected value of the joint probability;

[0040] The joint probability expectation value of two consecutive states is calculated iteratively until the preset stopping condition is met, and the optimized hidden Markov model is obtained.

[0041] Preferably, the emotion acquisition module includes a physiological detection module, a real-time emotion detection module and an emotion visualization module;

[0042] A physiological detection module is used to detect the value of emotion fluctuations and input the value of emotion fluctuations into the real-time emotion detection module;

[0043] A real-time emotion detection module is used to calculate the measurement coefficient according to the emotion fluctuation value and judge the emotion valence based on the measurement coefficient;

[0044] Emotion visualization module, used to express emotions and provide corresponding emotional feedback.

[0045] Preferably, the physiological detection module includes:

[0046] Collecting skin conductance signals, and sequentially performing gain amplification processing, filtering processing, and rectification and integration processing on the skin conductance signals;

[0047] Based on the skin conductance signal, the skin electrical change rate and skin electrical response are calculated respectively;

[0048] The calculation formula of skin electrical change rate is:

[0049] rSCR = lg|(average skin galvanic response - SCB) / SCB|;

[0050] The calculation formula of skin galvanic response is:

[0051] SCR=SCR max -SCR min ;

[0052] In the formula, rSCR represents the skin electrical change rate, SCB represents the skin electrical response;

[0053] The skin electrical change rate and skin electrical response are compared with the preset thresholds respectively. When the skin electrical change rate and skin electrical response both meet the preset thresholds, they are input into the real-time emotion detection module.

[0054] Preferably, the real-time emotion detection module calculates the measurement coefficient according to the emotion fluctuation value, and determines the emotion valence based on the measurement coefficient, including:

[0055] The laterality coefficient is calculated based on the galvanic skin response, and the emotional valence is judged based on the laterality coefficient;

[0056] The calculation formula of the side coefficient is:

[0057] P=×(SCR left -SCR right ) / (SCR left +SCR right );

[0058] Where p represents the side coefficient; SCR left SCR stands for the skin electrical response of the left hand; right represents the galvanic skin response of the right hand;

[0059] The emotional valence is judged based on the laterality coefficient; when the laterality coefficient is a negative value, it indicates a right-hand dominant reaction and the current emotion is a negative emotion; when the laterality coefficient is a positive value, it indicates a left-hand dominant reaction and the current emotion is a positive emotion.

[0060] According to a second aspect of the present invention, there is provided an intelligent emotion soothing method in a noisy environment, the intelligent emotion soothing method in a noisy environment comprising the following steps:

[0061] S1. Match tasks with high emotions and low emotions based on task complexity to generate high emotion tasks and low emotion tasks;

[0062] S2, collecting sound information in high-emotion tasks and classifying and processing the sound information;

[0063] S3, collect emotion types and emotion change trends in low emotion tasks and perform emotion recognition;

[0064] S4, combining the emotion recognition result and the sound information classification result to sequentially perform noise reduction and sound conversion processing on the ambient noise to achieve a soothing effect on negative emotions;

[0065] S5. Output the sound signal to the human ear and provide positive emotional guidance through the emotion soothing device.

[0066] According to a third aspect of the present invention, a smart emotion soothing device in a noisy environment is provided. The smart emotion soothing device in a noisy environment also includes an earphone component and a bracelet component; the earphone component is electrically connected to the bracelet component.

[0067] Preferably, the earphone assembly comprises an earphone body, a player is arranged on one side of the inner part of the earphone body, an ear hook is arranged on one side of the outer part of the earphone body, and a sound transmitter is arranged between the ear hook and the earphone body;

[0068] The other side of the earphone body is provided with an earphone switch, a volume up button, a volume down button and a radio microphone in sequence, and the bottom of the earphone body is provided with an earphone charging port.

[0069] Preferably, the bracelet assembly includes a bracelet body, a display screen is provided at one end of the middle part of the bracelet body, a charging indicator light is provided on one side of the top of the display screen, an emotion detection display light is provided on the other side of the top of the display screen, and a music play button is provided at the bottom of the display screen; a bracelet charging port is provided on one side of the outside of the bracelet body, and a bracelet switch is provided on the other side of the outside of the bracelet body; a watch strap is provided on the top of the bracelet body, an emotion identifier is provided on one end of the watch strap, and a magnetic absorber is provided on the other end of the watch strap;

[0070] A skin electrical activity detector is arranged at the other end of the middle part of the wristband body, and a temperature sensor is arranged at the bottom of the skin electrical activity detector.

[0071] The beneficial effects of the present invention are: the intelligent emotion soothing system provided by the present invention is suitable for auditory information processing and worker emotion soothing in a variety of working environments, and is highly practical. People who work in stressful and noisy environments for a long time can use it in different social environments, thereby effectively soothing the emotions of different operators according to their emotional needs, improving work experience, and soothing negative emotions. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0073] Figure 1 is a principle block diagram of an intelligent emotion soothing system in a noisy environment according to an embodiment of the present invention;

[0074] Figure 2 is a schematic diagram of an implementation of a task and emotion matching module in an intelligent emotion soothing system in a noisy environment according to an embodiment of the present invention;

[0075] Figure 3 It is a principle block diagram of a sound input module in an intelligent emotion soothing system in a noisy environment according to an embodiment of the present invention;

[0076] Figure 4 This is a principle block diagram of an emotion collection module in an intelligent emotion soothing system in a noisy environment according to an embodiment of the present invention;

[0077] Figure 5is a principle block diagram of an emotion soothing module in an intelligent emotion soothing system in a noisy environment according to an embodiment of the present invention;

[0078] Figure 6 It is a principle block diagram of a neural network model in an intelligent emotion soothing system in a noisy environment according to an embodiment of the present invention;

[0079] Figure 7 is a flow chart of an intelligent emotion soothing method in a noisy environment according to an embodiment of the present invention;

[0080] Figure 8 is a flow chart of establishing an emotion and sound matching library in an intelligent emotion soothing system in a noisy environment according to an embodiment of the present invention;

[0081] Fig. 9 is a structural schematic diagram of an earphone component in an intelligent emotion soothing device in a noisy environment according to an embodiment of the present invention;

[0082] Fig.10 is a schematic structural diagram of an earphone assembly of an intelligent emotion soothing device in a noisy environment from another angle according to an embodiment of the present invention;

[0083] Fig.11 is a structural schematic diagram of a wristband component in a smart emotion soothing device in a noisy environment according to an embodiment of the present invention;

[0084] Fig.12 is a front view of a wristband component in a smart emotion soothing device in a noisy environment according to an embodiment of the present invention;

[0085] Fig.13 is a rear view of a wristband component in a smart emotion soothing device in a noisy environment according to an embodiment of the present invention;

[0086] Fig.14 This is a schematic diagram of functions in a touch screen display in an intelligent emotion soothing system in a noisy environment according to an embodiment of the present invention;

[0087] Fig.15 is a circuit system structure diagram of an intelligent emotion soothing system in a noisy environment according to an embodiment of the present invention;

[0088] Fig.16 is a schematic diagram of the structure of an emotion recognizer in an intelligent emotion soothing device in a noisy environment according to an embodiment of the present invention;

[0089] Fig.17 1 is a schematic structural diagram from another angle of an emotion recognizer in an intelligent emotion soothing device in a noisy environment according to an embodiment of the present invention.

[0090] In the figure:

[0091] 1. Task and emotion matching module; 2. Sound input module; 201. Environmental sound acquisition module; 202. Sound screening and filtering module; 203. Sound recognition module; 204. Sound classification module; 3. Emotion acquisition module; 301. Physiological detection module; 302. Real-time emotion detection module; 303. Emotion visualization module; 4. Emotion soothing module; 401. Sound noise reduction module; 402. AI intelligent processing module; 403. Sound adjustment module; 404. Manual selection module; 405. Adaptive matching module; 5. Sound output module; 6. Headphone assembly; 601. Headphone body; 602. Player; 603. Ear hook; 604. Sound transmitter; 605. Headphone switch; 606. Volume up button; 607. Volume down button; 608. Radio microphone; 609. Headphone charging port; 7. Bracelet Components; 701, bracelet body; 702, display screen; 703, charging indicator light; 704, emotion detection display light; 705, music play button; 706, bracelet charging port; 707, bracelet switch; 708, strap; 709, emotion recognizer; 7091, bottom plate; 70911, expression sensing contact; 70912, NFC wireless sensing button; 70913, charging pile; 70914, base chassis; 7092, hand lever; 70921, surface touch layer; 70922, material fixing layer; 70923, spring; 70924, pressure sensing component; 70925, transition rod; 70926, pressure sensing lever; 70927, base; 70928, pressure sensor; 70929, chassis; 710, magnetic absorber; 711, skin electrical activity detector; 712, temperature sensor. DETAILED DESCRIPTION

[0092] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments. They can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in this field should be able to understand other possible implementation methods and advantages of the present invention.

[0093] According to an embodiment of the present invention, an intelligent emotion soothing system, method and device in a noisy environment are provided.

[0094] Figure 1 An embodiment of an intelligent emotion soothing system in a noisy environment of the present invention is shown.

[0095] In this optional embodiment, the intelligent emotion soothing system in a noisy environment includes a task and emotion matching module 1, a sound input module 2, an emotion collection module 3, an emotion soothing module 4 and a sound output module 5;

[0096] The task and emotion matching module 1 is used to match the task with high emotion and low emotion based on the task complexity, and generate high emotion tasks and low emotion tasks.

[0097] Among them, Figure 2 As shown, the task and emotion matching module 1 matches the task with high emotion and low emotion based on the task complexity to generate high emotion tasks and low emotion tasks, including:

[0098] Calculate the task complexity and compare it with the preset threshold. The calculation formula of the task complexity is:

[0099] Task Complexity Score(TCS)=∑(w i × i );

[0100] In the formula, w i represents the weight of the i-th factor; r i represents the quantitative score of the i-th factor; TCS represents task complexity.

[0101] According to the comparison between task complexity and preset threshold, task attributes are divided into high emotion tasks, medium emotion tasks and low emotion tasks.

[0102] The cognitive load of the medium-emotion task is measured, and the medium-emotion task is divided into high-emotion task and low-emotion task according to the comparison between the cognitive load measurement result and the preset threshold.

[0103] The high emotion tasks are respectively transmitted to the sound input module 2, and the low emotion tasks are respectively transmitted to the emotion collection module 3.

[0104] It should be noted that i considers three factors, namely the number of task steps (S), the types of skills required (K), and time pressure (T), with weights of 0.3, 0.4, and 0.3 respectively. The quantitative score of the number of task steps S is determined by the task steps; the quantitative score of the types of skills K is determined by the number of skills; and the quantitative score of time pressure T is determined by subjective assessment (high, medium, and low pressure).

[0105] It should be noted that if TCS≤2.6, the task complexity is relatively small and it is a high-emotion task; if 5.9>TCS>2.6, the task complexity is relatively large and it is a medium-emotion task; if TCS≥5.9, the task complexity is very large and it is a low-emotion task.

[0106] Secondly, the cognitive load of workers was measured when 5.9>TCS>2.6, and further emotional task screening was performed for the neutral emotional task. The heart rate variability (HRV) values ​​were measured for 2-5 minutes using a photoelectric heart rate sensor, and the variability of consecutive heartbeat intervals was evaluated using time domain analysis indicators such as RMSSD (square root of the root mean square deviation) and SDNN (mean of the standard deviation).

[0107] When RMSSD is (27±12) ms and SDNN is (141±39) ms, it indicates that the cognitive load is low and it is a high-emotion task; when RMSSD>(27±16) ms and SDNN>(141±45) ms, it indicates that the cognitive load is high and it is a low-emotion task.

[0108] The sound input module 2 is used to collect sound information in high-emotion tasks and classify and process the sound information.

[0109] It should be noted that if Figure 3 As shown, the sound input module 2 is mainly aimed at collecting, screening, identifying and classifying the sound information of the high-emotion task environment in the task and emotion matching module 1. After the sound input module 2 is started, the sound information of the surrounding environment is collected, screened, identified and classified. Collection refers to collecting sound information in the environment through the sound receiving microphone 608; screening refers to the intelligent chip extracting sounds above the threshold and filtering out sounds below 5dBa (decibel); identification refers to the identification of the types and sound elements (sound frequency, sound intensity, sound duration, sound speed) of the sound types (or type sets) above the threshold; classification refers to storing the sound materials in the corresponding sound library (human voice, machine sound, animal sound, natural sound).

[0110] The sound input module 2 includes an environmental sound collection module 201, a sound screening and filtering module 202, a sound recognition module 203 and a sound classification module 204;

[0111] The environmental sound collection module 201 is used to collect sound information in high-emotion tasks using a sound receiving microphone 608;

[0112] The sound screening and filtering module 202 is used to extract a set of sound types that meet a preset range from the sound information, and filter out the sound information outside the preset range.

[0113] It should be noted that the sound screening and filtering module 202 mainly extracts sounds above 5dBa (decibel) and filters out sounds below 5dBa (decibel) through a smart chip.

[0114] The sound recognition module 203 is used to sequentially recognize the sound types and sound elements of the extracted sound type set.

[0115] It should be noted that the sound recognition module 203 mainly recognizes and divides the sound type and sound elements. Among them, the recognition range of the sound frequency is divided into 7 frequency bands: extremely low frequency, low frequency, medium-low frequency, medium frequency, medium-high frequency, high frequency, and extremely high frequency. Extremely low frequency filtering (type: Bessel; mode: extremely low frequency; cut-off frequency 40 Hz, standardized to -0.1dB), low frequency filtering (type: Bessel; mode: low frequency; cut-off frequency 40 Hz, medium-low frequency cut-off 80 Hz, standardized to -0.1dB), medium-low frequency filtering (type: Bessel; mode: medium-low frequency; cut-off frequency 80 Hz, medium frequency cut-off 160 Hz, standardized to -0.1dB), medium frequency filtering (type: Bessel; mode: medium frequency; cut-off frequency 160 Hz, medium-high frequency cut-off 1280 Hz, standardized to -0.1dB B), mid-high frequency filtering (type: Bessel; mode: mid-high frequency; cut-off frequency 1280 Hz, high frequency cut-off 2560 Hz, standardized to -0.1dB), high frequency filtering (type: Bessel; mode: high frequency; cut-off frequency 2560 Hz, ultra-high frequency cut-off 5120 Hz, standardized to -0.1dB), ultra-high frequency filtering (type: Bessel; mode: ultra-high frequency; cut-off frequency 5120 Hz, standardized to -0.1dB), all sounds are saved in WAV format, with a sampling frequency of 44100 Hz stereo, 16 bits.

[0116] The recognition range of sound intensity is divided into five levels: micro intensity, low intensity, medium intensity, high intensity, and super high intensity. When it is less than 30dB SPL, it is micro intensity; when it is 30-55dB SPL, it is low intensity; when it is 55-70dB SPL, it is medium intensity; when it is 70-85dB SPL, it is high intensity; when it exceeds 85dB SPL, it is super high intensity.

[0117] The recognition range of sound speed is divided into slow, moderate and fast. Slow speed is based on the adagio beat, which is about 52 beats per minute; medium speed is based on the mezzo beat, which is 66 beats per minute; fast speed is based on the allegro beat, which is 132 beats per minute.

[0118] The sound classification module 204 is used to distinguish the sound features in the sound information using a neural network model.

[0119] It should be noted that the sound classification module 204 constructs an adaptive learning model based on a neural network to distinguish between human voice and machine voice. Among them, the characteristics of human voice are: the sound frequency range is between 85Hz and 255Hz, the sound intensity (loudness) is between 60dB and 80dB, the sound harmonic component is greater than 3, the resonance peak is usually between 100Hz and 500Hz, and the MFCC is a periodic change. The characteristics of machine sound are: a wider frequency range, a wider intensity range, a harmonic component is less than or equal to 3, and the MFCC (Mel Frequency Cepstral Coefficients) feature is a stable fluctuation.

[0120] Among them, the frequency components of the sound are obtained through Fourier transform, and the sound intensity is obtained by calculating the average energy or peak value of the sound signal; the timbre is obtained by resonance peak analysis; the resonance peak is determined by spectrum analysis to determine the resonance frequency of the sound; MFCC uses Mel-frequency cepstral coefficients to extract speech features.

[0121] As shown in 6, Python language and neural network model are used to extract sound features, design neural network and train the model. The code is as follows:

[0122]

[0123]

[0124] The sound classification module 204 includes the following steps when distinguishing the sound features in the sound information using the neural network model:

[0125] Use Fourier transform technology to obtain sound frequency components from sound information;

[0126] Calculate the average energy of the sound signal based on the sound frequency components to obtain the sound intensity; use frequency analysis technology to determine the sound resonance frequency to obtain the timbre characteristics; use Mel frequency cepstrum coefficients to extract speech features;

[0127] The sound intensity is converted into an observation symbol set, and the hidden Markov model is used to identify the clutter effect in the sound intensity and remove the clutter effect from the sound intensity.

[0128] It should be noted that the Hidden Markov Model is a statistical model that assumes that the state of the system cannot be directly observed (i.e., hidden state), but these states can be indirectly inferred through the observation sequence. In sound signal processing, these hidden states may represent different sound intensity levels or sound characteristics, while the observation sequence is the measurable data (such as sound intensity) directly extracted from the sound signal, specifically mapping the sound intensity to a finite set of observation symbols, each symbol representing a different sound intensity level.

[0129] The process of converting the sound intensity into an observation symbol set and using a hidden Markov model to identify the influence of clutter in the sound intensity includes the following steps:

[0130] Mapping the sound intensity into a finite set of observation symbols, wherein each observation symbol in the set of observation symbols represents a corresponding sound intensity level;

[0131] The observation symbols in the observation symbol set are combined into a propagation trajectory, and the propagation trajectory is used as an observation value sequence of the hidden Markov model;

[0132] The initial state transfer probability and initial observation probability are set for the hidden Markov model, and the parameters of the hidden Markov model are gradually adjusted using an iterative optimization algorithm to obtain an optimized hidden Markov model.

[0133] It should be noted that the continuous or high-dimensional sound data is converted into a discrete form that is more suitable for model processing, and the sequence of observation symbol sets (i.e., propagation trajectories) is used as the observation value sequence of the HMM (Hidden Markov Model), and the model starts with a known initial state transition probability and observation probability, and uses an iterative optimization algorithm to gradually adjust the model parameters. These parameter adjustments are based on the principle of maximizing the probability of occurrence of the observation sequence to reflect the statistical characteristics of real-world data. The trained and optimized HMM can effectively identify the influence of clutter in sound intensity.

[0134] Since HMM can capture the time series dependency in sound data (i.e., the dependency between previous and subsequent sound states), it can identify which changes are caused by clutter (such as background noise, interference, etc.), and once these clutters are identified, they can be removed from the analysis, leaving only the signal changes caused by actual sound events. The hidden Markov model not only improves the purification quality of sound signals, but also enhances the ability of subsequent neural network models to recognize the characteristics of human and machine sounds.

[0135] Among them, using an iterative optimization algorithm to gradually adjust the hidden Markov model parameters to obtain an optimized hidden Markov model includes the following steps:

[0136] Extracting the initial parameters of the hidden Markov model and calculating the forward probability and backward probability of the hidden Markov model at a preset time point, wherein the initial parameters are the initial state transition probability and the initial observation probability;

[0137] Combine the forward and backward probabilities to calculate the occupancy probability of a single state at any time point in the observation sequence and the joint probability of two consecutive states;

[0138] Calculate the expected value of the joint probability of the two consecutive states based on the joint probability of the two consecutive states, and update the initial parameters of the hidden Markov model based on the expected value of the joint probability;

[0139] The joint probability expectation value of two consecutive states is calculated iteratively until the preset stopping condition is met, and the optimized hidden Markov model is obtained.

[0140] The optimized hidden Markov model is used to identify the clutter effect in sound intensity.

[0141] The sound intensity, timbre characteristics and speech characteristics after eliminating the influence of noise are used as the input of the neural network model respectively, and the neural network model is used to distinguish the human voice characteristics and machine voice characteristics in the sound information.

[0142] It should be noted that the iterative optimization algorithm in the present invention is the Baum-Welch algorithm, which is a parameter estimation technique of the hidden Markov model, and iteratively adjusts the parameters of the model by maximizing the probability of observing the sequence data. This algorithm is used to process statistical models with hidden states, and by continuously optimizing the state transition probability and the observation probability, the prediction accuracy and data fit of the hidden Markov model are improved, and it is adapted to complex or hidden data structures, thereby effectively enhancing the performance and applicability of the hidden Markov model, making it a powerful tool for processing time series data and other complex data structures containing hidden states.

[0143] The emotion collection module 3 is used to collect emotion types and emotion change trends in low emotion tasks and perform emotion recognition.

[0144] The emotion acquisition module 3 includes a physiological detection module 301, a real-time emotion detection module 302 and an emotion visualization module 303;

[0145] The physiological detection module 301 is used to detect the emotion fluctuation value and input the emotion fluctuation value into the real-time emotion detection module.

[0146] The physiological detection module 301 includes the following steps when detecting the emotion fluctuation value and inputting the emotion fluctuation value into the real-time emotion detection module:

[0147] Collecting skin conductance signals, and sequentially performing gain amplification processing, filtering processing, and rectification and integration processing on the skin conductance signals;

[0148] Based on the skin conductance signal, the skin electrical change rate and skin electrical response are calculated respectively;

[0149] The skin electrical change rate and the skin electrical response are compared with the preset thresholds respectively. When the skin electrical change rate and the skin electrical response both meet the preset thresholds, they are input into the real-time emotion detection module 302.

[0150] It should be noted that if Figure 4As shown, the physiological detection module 301 is mainly used for the data collection and response of the arousal degree of personal emotions. The wristband component 7 is turned on by the wristband switch 707 to start the physiological detection module 301; the physiological detection module 301 inputs the detected emotional fluctuation of the wearer (reaching the preset value) into the real-time emotion detection module 302.

[0151] It should be noted that data acquisition and recording include: Ag / AgCl electrodes with a diameter of 10 mm, the electrodes are fed by a 0.5 V constant voltage transducer, and after signal acquisition, the signal is amplified by a gain of 10,000 times, filtered by a first-order high-pass filter of 16 Hz and a second-order low-pass filter of 340 Hz, and then rectified and integrated with a time constant of 100 ms. The skin conductance response (SCR) is calculated by subtracting the average value of the 2-s baseline before the start of sliding from the average value of the 2-s baseline after the start of sliding.

[0152] The skin conductance change rate (Rate of SCR, rSCR) and skin conductance response (SCR) levels are used as indicators for detecting emotional fluctuations. The calculation formula for the skin conductance change rate is:

[0153] rSCR = lg|(average skin galvanic response - SCB) / SCB|;

[0154] For the galvanic skin response, the maximum amplitude of the SCR during the information presentation phase was subtracted from the minimum amplitude. The calculation formula for the galvanic skin response is:

[0155] SCR=SCR max -SCR min ;

[0156] In the formula, rSCR represents the skin electrical change rate, SCB represents the skin electrical response;

[0157] It should be noted that the preset values ​​include: when |rSCR|≥0.6, SCR≥2.0 μmho (micro-ohm), the preset values ​​are reached, then the real-time emotion detection module 302 is started.

[0158] The real-time emotion detection module 302 is used to calculate the measurement coefficient according to the emotion fluctuation value, and determine the emotion valence based on the measurement coefficient.

[0159] It should be noted that the emotion recognition module includes a real-time emotion detection module 302 and an emotion visualization module 303. It is mainly used for collecting, analyzing and transmitting data such as the wearer's emotion valence and type, and sending the data results to the emotion soothing module 4.

[0160] The emotion recognition module in the bracelet component 7 (two for the left hand and the right hand) uses an Ag / AgCl electrode with a diameter of 10 mm to collect the skin conductance response (SCR) value, and then calculates the sideness coefficient to determine the emotional valence, and performs emotion recognition analysis on the smart chip in the bracelet component 7.

[0161] In the real-time emotion detection module 302, the wristband components 7 of the left and right hands respectively measure the skin conductance responses of the left and right hands. When detecting the skin conductance response, in order to reduce error noise, only SCR greater than 0.05 μmho is analyzed, and the lateral coefficient value is used to judge the emotional valence.

[0162] The real-time emotion detection module 302 calculates the measurement coefficient according to the emotion fluctuation value and determines the emotion valence based on the measurement coefficient, including:

[0163] The laterality coefficient is calculated based on the galvanic skin response, and the emotional valence is judged based on the laterality coefficient;

[0164] The calculation formula of the side coefficient is:

[0165] P = 100 × (SCR left -SCR right ) / (SCR left +SCR right );

[0166] Where p represents the side coefficient; SCR left SCR stands for the skin electrical response of the left hand; right Represents the electrical skin response of the right hand.

[0167] The emotional valence is judged based on the laterality coefficient; when the laterality coefficient is a negative value, it indicates a right-hand dominant reaction and the current emotion is a negative emotion; when the laterality coefficient is a positive value, it indicates a left-hand dominant reaction and the current emotion is a positive emotion.

[0168] The emotion visualization module 303 is used to express emotions and provide corresponding emotion feedback.

[0169] It should be noted that in the emotion visualization module 303, sound and light are used to allow people to quickly express their emotions and give corresponding feedback through physical buttons. Figure 16-Figure 17The emotion recognizer 709 on the wristband component 7 is composed of a bottom plate 7091 and four hand bars 7092. The bottom plate 7091 is provided with an NFC wireless sensing button 70912. Different hand bars 7092 have different patterns and touch patterns on their surfaces, representing different moods (joy, anger, sorrow, and happiness). The hand bar 7092 is placed on the bottom plate 7091 and pressed down. After being sensed, the bottom plate 7091 can emit a sound corresponding to the emotion. The surface of the hand bar 7092 representing an angry expression is a jagged three-dimensional protruding pattern. After pressing down, the bottom plate 7091 will emit an unpleasant sound of "Aiyaya", vividly combining sound with touch.

[0170] The bottom plate 7091 is composed of expression sensing contacts 70911, NFC wireless sensing buttons 70912, charging piles 70913 and base chassis 70914. The hand lever 7092 is composed of a surface touch layer 70921, a material fixing layer 70922, a spring 70923, a pressure sensing component 70924, a transition rod 70925, a pressure sensing lever 70926, a base 70927 (including a sound hole), a pressure sensor 70928, and a chassis 70929 (wireless charging contacts).

[0171] Use Arduino to set four buttons to control four small lights. The electronic components needed include: Arduino Uno development board, four operation buttons, four LED small lights, eight 220 ohm resistors and several jumpers. The four buttons (hand levers) control four color small lights respectively, that is, when the four hand levers 7092 are pressed, the hand levers 7092 flash different colors and emit corresponding sounds, and can present a gradual soft light with the strength of the pressure.

[0172] The following defines the pin numbers of four buttons and four LEDs. In the setup() function, the LED pins are set to output mode, and the button pins are set to input mode. The built-in pull-up resistors are enabled. When the buttons are not pressed, they will read as high (HIGH), and when the buttons are pressed, they will read as low (LOW). In the 1oop() function, the state of each button is read in a loop, and the corresponding LED is controlled according to the state. If the button state is low, the corresponding LED is lit; if the button state is high, the corresponding LED is extinguished. It should be noted that when actually connecting the hardware, one end of the button needs to be connected to the digital pin of the Arduino and the other end is connected to the ground (GND) to correctly implement the input function of the button. At the same time, the positive pole of the LED is connected to the digital pin of the Arduino, and the negative pole of the LED should be connected to the ground or connected to the ground through an appropriate current limiting resistor. The specific code and function implementation are shown in Table 1.

[0173] Table 1 Specific code and function implementation table

[0174]

[0175]

[0176]

[0177] The emotion soothing module 4 is used to perform noise reduction and sound conversion processing on the environmental noise in sequence according to the emotion recognition result and the sound information classification result, so as to achieve the soothing effect on negative emotions.

[0178] It should be noted that if Figure 5 As shown, the emotion soothing module 4 mainly soothes the workers' emotions by reducing the ambient noise and converting the sound, and is composed of a sound noise reduction module 401, an AI intelligent processing module 402, a sound adjustment module 403, a manual selection module 404 and an adaptive matching module 405.

[0179] The sound noise reduction module 401 is used to reduce or eliminate noise in the environment. External noise is monitored through built-in electronic devices and sensors, and sound waves with opposite phases are generated to offset or reduce the impact of external noise. This is usually achieved by using a miniature microphone to capture environmental noise and playing the corresponding anti-phase sound waves in headphones or speakers. The sound noise reduction range is the machine sound model range of the sound classification module.

[0180] The AI ​​intelligent processing module 402 is mainly used to generate environmental sounds (such as music, human voices, etc.) that can effectively ease the wearer's mood. The AI ​​intelligent processing module 402 distinguishes whether it is a positive emotion and performs different processing. When the sound of positive emotions is input into the AI ​​intelligent processing module 402, the AI ​​generates similar music, human voices (children's voices, jokes), machine sounds, animal sounds, and natural sounds, that is, relevant sounds that give the wearer a positive emotional orientation.

[0181] The sound adjustment module 403 is used for matching and storing the sound with positive emotions and negative emotions. The sound adjustment module 403 is composed of an adaptive matching module and a manual selection module.

[0182] The manual selection module 404 is used to perform option click operations on task types, sound modes, emotion types and settings according to the work situation and the wearer's personal preferences. Among them, the task type includes various work tasks set in the mobile phone before work. After entering the number of task steps (S), the type of skills required (K) and the time pressure (T), the intelligent processing chip completes the processing of the task and emotion matching module. The sound mode includes sound noise reduction, AI processing and music playback. The main operation is that the wearer can manually choose to turn it on or off. Emotion types include positive emotions and negative emotions. When the adaptive matching does not meet the wearer's psychological expectations, the wearer can manually select positive emotions or negative emotions. Settings include time, brightness and personalized wallpaper selection of the display.

[0183] like Figure 8 As shown, the adaptive matching module 405 is mainly used to establish an emotion and sound matching library, including a sound type, sound element and negative emotion matching library, and a sound type, sound element and positive emotion matching library. During the worker's work, firstly, according to the sound type and sound element value, the body temperature measurement value is extracted based on the MAX30205 body temperature sensor, and then the side coefficient is calculated based on the SCR, so as to establish a sound type, sound element and negative emotion matching library and a sound type, sound element and positive emotion matching library.

[0184] The sound output module 5 is used to output the sound signal to the human ear and provide positive emotional guidance through the emotion soothing device.

[0185] It should be noted that the sound output module 5 is used to generate sound signals and output them to human ears, and the relevant sounds with positive emotion guidance are played through the earphone component 6 to achieve the effect of alleviating emotions.

[0186] Sound playback mode: When the physiological detection module 301 detects that the wearer has strong emotional fluctuations (reaching the preset value), the relevant sound of positive emotion orientation is automatically played through the wearable device; when the emotional fluctuation detected by the physiological detection module 301 does not reach the preset value, the wearer can also press the bracelet button option to start manual playback and play the relevant sound of positive emotion orientation. The manual playback button is placed inside the bracelet component 7, including functions such as sound type selection, item switching within the type, volume adjustment, and environmental noise reduction.

[0187] Figure 7 An embodiment of an intelligent emotion soothing method in a noisy environment of the present invention is shown.

[0188] In this optional embodiment, the intelligent emotion soothing method in a noisy environment includes the following steps:

[0189] S1. Match tasks with high emotions and low emotions based on task complexity to generate high emotion tasks and low emotion tasks;

[0190] S2, collecting sound information in high-emotion tasks and classifying and processing the sound information;

[0191] S3, collect emotion types and emotion change trends in low emotion tasks and perform emotion recognition;

[0192] S4, combining the emotion recognition result and the sound information classification result to sequentially perform noise reduction and sound conversion processing on the ambient noise to achieve a soothing effect on negative emotions;

[0193] S5. Output the sound signal to the human ear and provide positive emotional guidance through the emotion soothing device.

[0194] Figure 3-4 An embodiment of an intelligent emotion soothing device in a noisy environment of the present invention is shown.

[0195] In this optional embodiment, the intelligent emotion soothing device in a noisy environment also includes an earphone component 6 and a bracelet component 7; the earphone component 6 is electrically connected to the bracelet component 7.

[0196] Among them, the earphone component 6 includes an earphone body 601, a player 602 is arranged on the inner side of the earphone body 601, an ear hook 603 is arranged on the outer side of the earphone body 601, and a sound transmitter 604 is arranged between the ear hook 603 and the earphone body 601; an earphone switch 605, a volume increase button 606, a volume decrease button 607 and a radio microphone 608 are arranged in sequence on the other side of the outer side of the earphone body 601, and an earphone charging port 609 is arranged at the bottom of the earphone body 601.

[0197] It should be noted that the sound transmitter 604 is arranged between the ear hook 603 and the earphone body 601, and serves as an overall connection; the player 602 is located inside the earphone body 601, and serves as a music player; the earphone switch 605, the volume up button 606, the volume down button 607, and the radio microphone 608 are all located on the earphone component 6, with a simple design, which is convenient for the wearer to distinguish and use; the earphone charging port 609 is located at the bottom of the smart earphone body 601, and is used for device charging; the ear hook 603 fits the back of the ear, so that the wearer can wear it for a long time without hurting the ear; the radio microphone 608 is located on the outside of the music playback device, which is convenient for collecting ambient sound and is not affected by other factors, so as to collect task information.

[0198] The adaptive matching module 405 and the AI ​​intelligent processing module 402 are both located in the earphone component 6, which is convenient for collecting and organizing the collected environmental information and starting the AI ​​intelligent processing module 402, and finally calling the player 602 to play music.

[0199] Among them, the bracelet component 7 includes a bracelet body 701, a display screen 702 is arranged at one end of the middle part of the bracelet body 701, a charging indicator light 703 is arranged on one side of the top of the display screen 702, an emotion detection display light 704 is arranged on the other side of the top of the display screen 702, and a music play button 705 is arranged at the bottom of the display screen 702; a bracelet charging port 706 is arranged on one side of the outside of the bracelet body 701, and a bracelet switch 707 is arranged on the other side of the outside of the bracelet body 701; a strap 708 is arranged on the top of the bracelet body 701, an emotion identifier 709 is arranged at one end of the strap 708, and a magnetic absorber 710 is arranged at the other end of the strap 708; a skin electrode activity detector 711 is arranged at the other end of the middle part of the bracelet body 701, and a temperature sensor 712 is arranged at the bottom of the skin electrode activity detector 711.

[0200] It should be noted that the functions of the various structures of the wristband component 7 are as follows:

[0201] The charging indicator light 703 is continuously red when charging, has no display light in standby mode, and is green when in use and when charging is complete.

[0202] The emotion detection display light 704 is off when the wearer is in a positive or calm emotion stage, and lights up blue in a negative emotion stage, serving as a reminder.

[0203] The music play button 705 can directly call the wearer to assist the music play device to play music, which is convenient for the wearer to use.

[0204] The magnetic absorber 710 is at the top of the strap and is used to fix the bracelet assembly 7.

[0205] The skin conductance detector 711 is located on the strap 708 and has a built-in Ag / AgCl electrode with a diameter of 10 mm to collect skin conductance response (SCR) values.

[0206] The music play button 705 can directly call the earphone component 6 to play random music;

[0207] The temperature sensor 712 has a built-in MAX30205 body temperature sensor to extract body temperature measurement values.

[0208] Display screen 702 (touch screen) enters the function display page, which includes four functions: task selection, sound mode, emotion selection and setting. After the worker touches the screen and enters the task selection page, he slides the screen up and down to select a task (click to confirm the selection plan), and the intelligent processing chip in the bracelet component 7 will process the task and emotion matching module 1. After sliding the screen left and right to enter the sound mode function, slide the screen up and down to enter the sub-function page of sound noise reduction, AI processing, and music playback. The wearer can manually select the sub-function to be on / off. After sliding the screen left and right to enter the emotion type function, slide the screen up and down to enter the positive emotion or negative emotion sub-function page. The wearer can manually select the positive emotion or negative emotion function that meets the wearer's psychological expectations. After sliding the screen left and right to enter the setting function, slide the screen up and down to enter the sub-function page of personalized time, brightness, and wallpaper of the display screen for setting. The user interface function display on the display screen 702 is as follows: Fig.14 shown.

[0209] The built-in chip of the wristband component 7 is used for task and emotion matching module 1 analysis, skin electrical activity detector 711 and temperature sensor 712 to collect data for intelligent analysis, and connects the sound output module 5 with the earphone component 6 to control the sound noise reduction and music playback in the earphone component 6. The circuit system structure diagram is shown in FIG. Fig.15 shown.

[0210] To sum up, with the help of the above-mentioned technical scheme of the present invention, the intelligent emotion soothing system provided by the present invention is suitable for auditory information processing and worker emotion soothing in a variety of working environments. It is highly practical and can be used by people who work in stressful and noisy environments for a long time in different social environments, thereby effectively performing emotion soothing processing according to the emotional needs of different operators, improving work experience, and soothing negative emotions.

[0211] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent emotion soothing system in a noisy environment, characterized in that: The intelligent emotion soothing system in a noisy environment comprises a task and emotion matching module (1), a sound input module (2), an emotion collection module (3), an emotion soothing module (4) and a sound output module (5); The task and emotion matching module (1) is used to match the task with high emotion and low emotion based on the task complexity, and generate a high emotion task and a low emotion task; The sound input module (2) is used to collect sound information in high-emotion tasks and classify and process the sound information; The emotion collection module (3) is used to collect emotion types and emotion change trends in low emotion tasks and perform emotion recognition; The emotion soothing module (4) is used to perform noise reduction and sound conversion processing on the environmental noise in sequence in combination with the emotion recognition result and the sound information classification result, so as to achieve a soothing effect on negative emotions; The sound output module (5) is used to output the sound signal to the human ear and provide positive emotional guidance through the emotion soothing device.

2. The intelligent emotion soothing system in a noisy environment according to claim 1, characterized in that: The task and emotion matching module (1) matches the task with high emotion and low emotion based on the task complexity to generate a high emotion task and a low emotion task, including: Calculate the task complexity and compare the task complexity with a preset threshold; The calculation formula of task complexity is: Task Complexity Score(TCS)=∑(w i ×r i ); In the formula, w i represents the weight of the i-th factor; r i represents the quantitative score of the i-th factor; TCS represents task complexity; According to the comparison between task complexity and preset threshold, the task attributes are divided into high emotion tasks, medium emotion tasks and low emotion tasks; The cognitive load of the medium-emotion task is measured, and the medium-emotion task is divided into a high-emotion task and a low-emotion task according to the comparison result of the cognitive load measurement result with the preset threshold; The high emotion tasks are respectively transmitted to the sound input module (2), and the low emotion tasks are respectively transmitted to the emotion collection module (3).

3. The intelligent emotion soothing system in a noisy environment according to claim 2, characterized in that: The sound input module (2) comprises an environmental sound collection module (201), a sound screening and filtering module (202), a sound recognition module (203) and a sound classification module (204); The environmental sound collection module (201) is used to collect sound information in high-emotion tasks using a sound receiving device; The sound screening and filtering module (202) is used to extract a set of sound types that meet a preset range from the sound information, and filter out the sound information outside the preset range; The sound recognition module (203) is used to sequentially recognize the sound types and sound elements of the extracted sound type set; The sound classification module (204) is used to distinguish the sound features in the sound information using a neural network model.

4. The intelligent emotion soothing system in a noisy environment according to claim 3, characterized in that: The sound classification module (204) comprises, when using a neural network model to distinguish sound features in sound information: Use Fourier transform technology to obtain sound frequency components from sound information; Calculate the average energy of the sound signal based on the sound frequency components to obtain the sound intensity; use frequency analysis technology to determine the sound resonance frequency to obtain the timbre characteristics; use Mel frequency cepstrum coefficients to extract speech features; The sound intensity is converted into an observation symbol set, and the clutter effect in the sound intensity is identified by using a hidden Markov model, and the clutter effect is removed from the sound intensity; The sound intensity, timbre characteristics and speech characteristics after eliminating the influence of noise are used as the input of the neural network model respectively, and the neural network model is used to distinguish the human voice characteristics and machine voice characteristics in the sound information.

5. The intelligent emotion soothing system in a noisy environment according to claim 4, characterized in that: The converting of the sound intensity into an observation symbol set and using a hidden Markov model to identify the influence of clutter in the sound intensity comprises the following steps: Mapping the sound intensity into a finite set of observation symbols, wherein each observation symbol in the set of observation symbols represents a corresponding sound intensity level; The observation symbols in the observation symbol set are combined into a propagation trajectory, and the propagation trajectory is used as an observation value sequence of the hidden Markov model; The initial state transfer probability and initial observation probability are set for the hidden Markov model, and the parameters of the hidden Markov model are gradually adjusted using an iterative optimization algorithm to obtain an optimized hidden Markov model; The optimized hidden Markov model is used to identify the clutter effect in sound intensity.

6. The intelligent emotion soothing system in a noisy environment according to claim 5, characterized in that: The method of gradually adjusting the parameters of the hidden Markov model using an iterative optimization algorithm to obtain an optimized hidden Markov model comprises the following steps: Extracting initial parameters of the hidden Markov model, and calculating the forward probability and backward probability of the hidden Markov model at a preset time point, wherein the initial parameters are the initial state transition probability and the initial observation probability; Combine the forward and backward probabilities to calculate the occupancy probability of a single state at any time point in the observation sequence and the joint probability of two consecutive states; Calculate the expected value of the joint probability of the two consecutive states based on the joint probability of the two consecutive states, and update the initial parameters of the hidden Markov model based on the expected value of the joint probability; The joint probability expectation value of two consecutive states is calculated iteratively until the preset stopping condition is met, and the optimized hidden Markov model is obtained.

7. The intelligent emotion soothing system in a noisy environment according to claim 6, characterized in that: The emotion collection module (3) comprises a physiological detection module (301), a real-time emotion detection module (302) and an emotion visualization module (303); The physiological detection module (301) is used to detect the emotion fluctuation value and input the emotion fluctuation value into the real-time emotion detection module; The real-time emotion detection module (302) is used to calculate the measurement coefficient according to the emotion fluctuation value, and judge the emotion valence based on the measurement coefficient; The emotion visualization module (303) is used to express emotions and provide corresponding emotion feedback.

8. The intelligent emotion soothing system in a noisy environment according to claim 7, characterized in that: The physiological detection module (301) includes, when detecting the emotion fluctuation value and inputting the emotion fluctuation value into the real-time emotion detection module: Collecting skin conductance signals, and sequentially performing gain amplification processing, filtering processing, and rectification and integration processing on the skin conductance signals; Based on the skin conductance signal, the skin electrical change rate and skin electrical response are calculated respectively; The calculation formula of skin electrical change rate is: rSCR = lg|(average skin galvanic response - SCB) / SCB|; The calculation formula of skin galvanic response is: SCR=SCR max -SCR min ; In the formula, rSCR represents the skin electrical change rate, SCB represents the skin electrical response; The skin electrical change rate and skin electrical response are compared with the preset thresholds respectively. When the skin electrical change rate and skin electrical response both meet the preset thresholds, they are input into the real-time emotion detection module.

9. The intelligent emotion soothing system in a noisy environment according to claim 8, characterized in that: The real-time emotion detection module (302) calculates the measurement coefficient according to the emotion fluctuation value, and determines the emotion valence based on the measurement coefficient, including: The laterality coefficient is calculated based on the galvanic skin response, and the emotional valence is judged based on the laterality coefficient; The calculation formula of the sideness coefficient is: P=100×(SCR left -SCR right ) / (SCR left +SCR right ); Where p represents the side coefficient; SCR left SCR stands for the skin electrical response of the left hand; right represents the galvanic skin response of the right hand; The emotional valence is judged based on the laterality coefficient; when the laterality coefficient is a negative value, it indicates a right-hand dominant reaction and the current emotion is a negative emotion; when the laterality coefficient is a positive value, it indicates a left-hand dominant reaction and the current emotion is a positive emotion.

10. An intelligent emotion soothing method in a noisy environment, used to implement the operation of the intelligent emotion soothing system in a noisy environment according to any one of claims 1 to 9, characterized in that: The intelligent emotion soothing method in a noisy environment comprises the following steps: S1. Match tasks with high emotions and low emotions based on task complexity to generate high emotion tasks and low emotion tasks; S2, collecting sound information in high-emotion tasks and classifying and processing the sound information; S3, collect emotion types and emotion change trends in low emotion tasks and perform emotion recognition; S4, combining the emotion recognition result and the sound information classification result to sequentially perform noise reduction and sound conversion processing on the ambient noise to achieve a soothing effect on negative emotions; S5. Output the sound signal to the human ear and provide positive emotional guidance through the emotion soothing device.

11. An intelligent emotion soothing device in a noisy environment, comprising an intelligent emotion soothing system in a noisy environment as claimed in any one of claims 1 to 9, characterized in that: The intelligent emotion soothing device in a noisy environment also includes an earphone component (6) and a bracelet component (7); Wherein, the earphone component (6) is electrically connected to the wristband component (7).

12. The intelligent emotion soothing device in a noisy environment according to claim 11, characterized in that: The earphone assembly (6) comprises an earphone body (601), a player (602) is arranged on one side of the interior of the earphone body (601), an ear hook (603) is arranged on one side of the exterior of the earphone body (601), and a sound transmitter (604) is arranged between the ear hook (603) and the earphone body (601); The other side of the outside of the earphone body (601) is provided with an earphone switch (605), a volume increase key (606), a volume decrease key (607) and a microphone (608) in sequence, and the bottom of the earphone body (601) is provided with an earphone charging port (609).

13. The intelligent emotion soothing device in a noisy environment according to claim 12, characterized in that: The wristband assembly (7) comprises a wristband body (701), a display screen (702) is arranged at one end of the middle part of the wristband body (701), a charging indicator light (703) is arranged on one side of the top of the display screen (702), an emotion detection display light (704) is arranged on the other side of the top of the display screen (702), and a music play button (705) is arranged at the bottom of the display screen (702); a wristband charging port (706) is arranged on one side of the outside of the wristband body (701), and a wristband switch (707) is arranged on the other side of the outside of the wristband body (701); a watchband (708) is arranged on the top of the wristband body (701), an emotion identifier (709) is arranged on one end of the watchband (708), and a magnetic absorber (710) is arranged on the other end of the watchband (708); A skin electrical activity detector (711) is arranged at the other end of the middle part of the bracelet body (701), and a temperature sensor (712) is arranged at the bottom of the skin electrical activity detector (711).

Citation Information

Cited By

  • Voice emotion recognition method and device for disposition of unhealthy assets

    CN119993217A