A Dynamic Ergonomics Optimization Method and System Based on Multimodal Sensing and Biofeedback

Abstract

This invention discloses a dynamic ergonomic optimization method and system based on multimodal sensing and biofeedback. The dynamic ergonomic optimization system based on multimodal sensing and biofeedback includes: a multimodal sensing module integrating a pressure sensor, an electromyography (EMG) sensor, and an environmental temperature and humidity sensor. The pressure sensor is suitable for real-time acquisition of head pressure distribution data of the target object. This application uses the multimodal sensing module to collect head pressure distribution data, neck muscle electrophysiological signals, and environmental temperature and humidity parameters in real time. Combined with the quantitative identification of fatigue state by the biofeedback module, the optimization decision module dynamically generates structural parameter adjustment instructions, so that parameters such as ventilation volume, padding hardness, and wearing tightness of the equipment adapt in real time to changes in physiological state and environment. This effectively alleviates muscle fatigue and discomfort after prolonged wear, and significantly improves the operational safety and sustainability of pilots, astronauts, and other long-term workers.

Description

Technical Field

[0001] This invention relates to the field of wearable device technology, specifically to a dynamic ergonomic optimization method and system based on multimodal perception and biofeedback. Background Technology

[0002] Ergonomic optimization technology plays a crucial role in wearable equipment such as fighter pilot helmets, spacesuit helmets, and industrial safety helmets. Its core objective is to improve the comfort, safety, and ease of use of the equipment by adapting it to human physiological characteristics and usage scenarios. Existing ergonomic optimization solutions are mainly based on static ergonomic evaluation systems. They design adaptation structures by pre-setting human body size data and single physical parameters (such as head circumference and baseline pressure threshold). Some solutions introduce pressure sensors or electromyography (EMG) sensors to monitor single-dimensional physiological signals. For example, pressure sensors collect head pressure distribution data, or EMG sensors monitor neck muscle fatigue, thereby providing fixed recommendations for adjusting adaptation parameters. In terms of feedback and interaction, existing technologies mostly rely on post-event data statistical analysis or simple audio-visual prompts. Some products support manual static adjustment of structural parameters such as ventilation volume and padding position, but a real-time dynamic optimization loop has not been formed.

[0003] However, existing technologies have significant drawbacks: First, the static optimization mode has prominent limitations. Traditional solutions complete the adaptation parameter settings before the equipment is worn, and cannot make real-time adjustments based on dynamic factors such as changes in human muscle fatigue and fluctuations in ambient temperature and humidity during use. This leads to accumulated muscle fatigue and decreased comfort after prolonged wear, which may affect operational safety, especially for people who need to work for long periods, such as fighter pilots and astronauts. Second, the ability to coordinate multi-factor optimization is lacking. Existing technologies mostly rely on single types of physiological signals or environmental parameters for optimization, failing to achieve deep integration of human physiological data (such as head pressure distribution and neck muscle electrophysiological signals) with environmental parameters (such as temperature and humidity). The existing framework is driven by a single factor, making it difficult to cope with the multivariate influences under complex working conditions. Third, the interaction and adaptation efficiency is insufficient. Existing interaction methods are mostly limited to single channels such as buttons and touch controls, which are inconvenient to operate when wearing heavy equipment (such as spacesuit helmets and fighter pilot helmets). Furthermore, adaptation solutions are mostly based on standardized data or simple algorithms, failing to incorporate historical usage data and deep learning technology, making it difficult to create personalized solutions that fit individual physiological characteristics. Fourth, the feedback mechanism lacks intuitiveness and real-time performance. Feedback on the adjustment effects of existing technologies is mostly delayed data analysis results, preventing users from knowing in real time the impact of parameter adjustments on ergonomic performance. This results in an incomplete optimization loop and makes it difficult to continuously improve adaptation accuracy.

[0004] Therefore, there is an urgent need for an ergonomic optimization method and system that breaks through the traditional static ergonomic evaluation model, can achieve multi-factor synergistic driving of physiological and environmental factors, support convenient multi-channel interaction, generate personalized adaptation solutions based on deep learning, and complete the dynamic optimization closed loop through real-time visual feedback. This would address the shortcomings of existing technologies in terms of dynamic adaptability, adaptation accuracy, ease of operation, and intuitive feedback, and meet the high requirements for long-term wearing comfort and operational safety in scenarios such as fighter pilot helmets, spacesuit helmets, and industrial safety helmets. Summary of the Invention

[0005] The purpose of this invention is to provide a dynamic ergonomic optimization method and system based on multimodal perception and biofeedback, so as to solve the problems existing in the prior art mentioned in the background.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a dynamic ergonomic optimization system based on multimodal sensing and biofeedback, comprising: The multimodal sensing module integrates a pressure sensor, an electromyography (EMG) sensor, and an environmental temperature and humidity sensor. The pressure sensor is suitable for real-time acquisition of head pressure distribution data of the target object, the EMG sensor is suitable for real-time acquisition of neck muscle electrophysiological signals of the target object, and the environmental temperature and humidity sensor is suitable for real-time acquisition of environmental temperature and humidity parameters of the environment in which the target object is located. A biofeedback module is communicatively connected to the multimodal sensing module. The biofeedback module is adapted to receive the electrophysiological signals of the neck muscles and the head pressure distribution data, identify the fatigue state of the target object through physiological signal analysis, and generate structural parameter adjustment instructions. The optimization decision module is communicatively connected to the multimodal perception module and the biofeedback module. The optimization decision module is suitable for calling deep learning algorithms, integrating real-time collected environmental temperature and humidity parameters, fatigue state information and historical usage data, and generating a personalized ergonomic adaptation parameter set in combination with ergonomic standards. The AR feedback module is communicatively connected to the optimization decision module. The AR feedback module is suitable for displaying the effect of structural parameter adjustment and the adaptation status of personalized ergonomic adaptation parameter set in real time. A cross-modal control module is adapted to receive system function commands triggered by a target object through a multi-channel operation mode, and the cross-modal control module is adapted to perform manual interactive control of system parameters and verification and calibration of adaptation parameters; The data storage module is suitable for storing various types of data collected in real time, fatigue state identification results, structural parameter adjustment records, personalized ergonomic adaptation parameter sets, and historical usage data. The data storage module provides data support for the model iterative optimization of deep learning algorithms and the dynamic updating of personalized ergonomic adaptation parameter sets.

[0007] Preferably, the biofeedback module collects the electrophysiological signals of the neck muscles of the target object through a brain-computer interface or a high-precision physiological signal monitoring unit. The biofeedback module uses feature engineering processing methods to extract the temporal and frequency domain features of the electromyographic signals. Furthermore, the biofeedback module completes the quantitative identification of fatigue state based on a preset multi-dimensional threshold system. The multi-dimensional threshold system includes the continuous wearing time threshold of the target object, the quantitative index of neck muscle fatigue, and the head pressure distribution uniformity threshold.

[0008] Preferably, the structural parameters corresponding to the structural parameter adjustment instructions include the ventilation volume of the wearable equipment, the hardness of the padding, the fit of the padding, the position of the head support point, and the tightness of the wear. In the structural parameter adjustment instructions output by the biofeedback module, the adjustment range of each structural parameter forms a dynamic matching relationship with the fatigue state quantification level.

[0009] Preferably, the historical usage data includes baseline data of the target object's physiological characteristics, dataset of changes in head pressure distribution, dynamic curve of neck muscle fatigue, historical records of structural parameter adjustments, environmental temperature and humidity adaptation data, iterative records of personalized ergonomic adaptation parameter sets, and feedback data on ergonomic optimization effects. The data storage module stores all historical usage data in a structured format.

[0010] Preferably, the personalized ergonomic adaptation parameter set includes the size specifications of the wearable equipment, the layout of the ventilation openings, the selection of the padding material, the design of the head support structure, the weight distribution optimization parameters, and the planning of the pressure dispersion area.

[0011] Preferably, the multi-channel operation modes supported by the cross-modal control module include voice command recognition, gesture motion capture, and eye-tracking.

[0012] Preferably, the wearable equipment includes, but is not limited to, fighter pilot helmets, spacesuit helmets, or industrial safety helmets.

[0013] The dynamic ergonomic optimization method based on any of the aforementioned systems includes the following steps: S1) The system’s multimodal sensing module is activated, which collects head pressure distribution data, neck muscle electrophysiological signals and temperature and humidity parameters of the target object’s environment in real time through integrated sensors, and transmits the collected data synchronously to the biofeedback module, optimization decision module and data storage module. S2) The biofeedback module performs noise filtering, signal amplification and feature normalization preprocessing on the received neck muscle electrophysiological signals and head pressure distribution data. It identifies the fatigue state by combining the continuous use time of the target object, the quantitative index of neck muscle fatigue and the head pressure distribution uniformity threshold, generates the corresponding structural parameter adjustment instructions and sends them to the optimization decision module. S3) The optimization decision module calls a deep learning algorithm pre-trained with historical usage data, integrates real-time environmental temperature and humidity parameters, fatigue status information and historical usage data in the data storage module, and generates a personalized ergonomic adaptation parameter set containing multiple adaptation parameters in accordance with ergonomic standards. S4) The AR feedback module provides real-time feedback on the structural parameter adjustment effect and the adaptation status of the personalized ergonomic adaptation parameter set to the target object in a visual form. S5) The target object triggers a manual adjustment command through the multi-channel operation mode of the cross-modal control module. After receiving the manual adjustment command, the optimization decision module updates the personalized ergonomic adaptation parameter set and simultaneously stores the updated personalized ergonomic adaptation parameter set to the data storage module, forming a dynamic optimization closed loop.

[0014] Preferably, the biofeedback module can determine fatigue status based on one of the neck muscle electrophysiological signals and head pressure distribution data, or by combining the two types of data. If based solely on the electrophysiological signals of the neck muscles, the focus is on analyzing the coefficient of variation of signal amplitude and the proportion of spectral energy. If the two types of data are combined, the head pressure distribution uniformity index and the time-domain characteristics of electromyography signals are integrated. The fatigue levels defined by the biofeedback module correspond to preset structural parameter adjustment threshold ranges, and the adjustment thresholds can be dynamically calibrated based on the baseline data of the target object's physiological characteristics. The fatigue state identification thresholds corresponding to different judgment methods can be calibrated independently.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1) This application uses a multimodal sensing module to collect head pressure distribution data, neck muscle electrophysiological signals and environmental temperature and humidity parameters in real time. Combined with the quantitative identification of fatigue state by the biofeedback module, the optimization decision module dynamically generates structural parameter adjustment instructions, so that the ventilation volume, padding hardness and wearing tightness of the equipment can be adapted in real time with changes in physiological state and environment. This effectively relieves muscle fatigue and discomfort after long-term wear, significantly improves the operational safety and continuity of pilots, astronauts and other long-term workers, and solves the core defect that traditional static adaptation solutions cannot cope with dynamic usage scenarios. 2) This application innovatively integrates human physiological data and environmental parameters to establish a multi-factor driven ergonomic optimization model. Through deep learning algorithms, it explores the correlation between head pressure distribution, neck muscle fatigue and environmental temperature and humidity. The generated personalized ergonomic adaptation parameter set covers multiple dimensions such as size specifications, ventilation layout, and weight distribution optimization. Compared with the existing single-factor driven optimization schemes, it is more targeted and can accurately match the usage needs under complex working conditions, greatly improving the compatibility between equipment and target objects. 3) This application supports multiple operation modes such as voice command recognition, gesture motion capture, and eye-tracking, without relying on traditional buttons or touch interaction. It is especially suitable for scenarios such as fighter pilots and astronauts who wear heavy equipment and have limited hand operation, reducing operation complexity and response time. At the same time, the cross-modal control module supports manual interaction control and adaptation parameter verification and calibration, taking into account the intelligence of automatic optimization and the flexibility of manual intervention, further improving the convenience and reliability of system operation. 4) This application is based on deep learning algorithms and integrates multi-source data such as the target object's physiological characteristic baseline data, historical usage records, and structural parameter adjustment ledgers. Combined with national ergonomic standards, it generates personalized adaptation solutions that can accurately match the physiological characteristic differences of different individuals. It is applicable to the personalized customization needs of multiple scenarios such as fighter pilot helmets, spacesuit helmets, and industrial safety helmets. It solves the problem that traditional standardized adaptation solutions cannot meet the needs of individual differences and significantly broadens the application scope of ergonomic optimization technology. 5) This application uses an AR feedback module to visualize the effect of structural parameter adjustment and adaptation status in real time, so that the target object can intuitively know the optimization effect. Combined with the continuous accumulation of real-time data and optimization records by the data storage module, it provides data support for the iterative optimization of deep learning models and the dynamic updating of adaptation schemes, forming a closed-loop optimization system of data collection, status recognition, scheme optimization, real-time feedback and iterative upgrade, promoting the continuous improvement of ergonomic optimization accuracy and adaptation performance, and ensuring the stability of ergonomic performance during long-term use of equipment. Attached Figure Description

[0016] Figure 1 This is a diagram of the optimized system framework of this application; Figure 2 This is a flowchart of the optimization method in this application. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] In the description of the invention, it should be noted that the terms "upper," "lower," "inner," "outer," "front end," "rear end," "both ends," "one end," and "the other end," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0019] In the description of the invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installed," "equipped with," "connected," etc., should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0020] In the description of the invention, it should be noted that the execution order of the steps is not limited by the sequence number. The possible changes in the order of some steps, the synchronous execution of steps, and the split execution of steps are all within the scope of protection of this application.

[0021] Please see Figure 1-2 This invention provides a technical solution: a dynamic ergonomic optimization system based on multimodal sensing and biofeedback, characterized in that it includes: The multimodal sensing module integrates a pressure sensor, an electromyography (EMG) sensor, and an environmental temperature and humidity sensor. The pressure sensor is suitable for real-time acquisition of head pressure distribution data of the target object, the EMG sensor is suitable for real-time acquisition of neck muscle electrophysiological signals of the target object, and the environmental temperature and humidity sensor is suitable for real-time acquisition of environmental temperature and humidity parameters of the environment in which the target object is located. The biofeedback module is connected in communication with the multimodal sensing module. The biofeedback module is suitable for receiving electrophysiological signals of the neck muscles and head pressure distribution data. It analyzes and identifies the fatigue state of the target object through physiological signal analysis and generates structural parameter adjustment instructions. The optimization decision module is connected to the multimodal perception module and the biofeedback module respectively. The optimization decision module is suitable for calling deep learning algorithms, integrating real-time collected environmental temperature and humidity parameters, fatigue status information and historical usage data, and generating a personalized ergonomic adaptation parameter set in combination with ergonomic standards. The AR feedback module communicates with the optimization decision module. The AR feedback module is suitable for displaying the effect of structural parameter adjustment and the adaptation status of personalized ergonomic adaptation parameter set in real time. The cross-modal control module is suitable for receiving system function commands triggered by the target object through multi-channel operation, and is also suitable for performing manual interactive adjustment of system parameters and verification and calibration of adaptation parameters. The data storage module is suitable for storing various types of data collected in real time, fatigue state identification results, structural parameter adjustment records, personalized ergonomic adaptation parameter sets, and historical usage data. The data storage module provides data empowerment for the model iterative optimization of deep learning algorithms and the dynamic updating of personalized ergonomic adaptation parameter sets.

[0022] Specifically, the dynamic ergonomic optimization system based on multimodal perception and biofeedback in this application achieves comprehensive real-time acquisition of head pressure distribution data, neck muscle electrophysiological signals, and environmental temperature and humidity parameters through a multimodal perception module integrating multiple types of sensors, providing rich and accurate basic data support for subsequent optimization. The communication connection design between the biofeedback module and the multimodal perception module ensures efficient transmission of physiological signals and timely identification of fatigue states, providing a reliable basis for structural parameter adjustment. The optimization decision module integrates multi-source data and calls deep learning algorithms, breaking the limitations of single data-driven approaches, and generating a personalized ergonomic adaptation parameter set that better suits individual needs and environmental characteristics. The AR feedback module enables real-time visualization of the adjustment effect, allowing the target object to intuitively grasp the adaptation state and improve the user experience. The cross-modal control module and data storage module respectively ensure the convenience of operation and the continuous reuse of data, laying the foundation for model iterative optimization.

[0023] Specifically, the entire system architecture of this application, from data collection, identification, decision-making, feedback, and storage, fundamentally breaks through the limitations of traditional static ergonomic optimization, realizes the synergistic driving of physiological and environmental factors, can dynamically respond to various changes during use, effectively improve the comfort, safety, and fitting accuracy of the equipment, and fully meet the high requirements of scenarios such as fighter pilot helmets, spacesuit helmets, and industrial safety helmets.

[0024] The biofeedback module collects the electrophysiological signals of the neck muscles of the target object through a brain-computer interface or a high-precision physiological signal monitoring unit. The biofeedback module uses feature engineering processing methods to extract the temporal and frequency domain features of the electromyographic signals. Furthermore, the biofeedback module completes the quantitative identification of fatigue state based on a preset multi-dimensional threshold system. The multi-dimensional threshold system includes the continuous wearing time threshold of the target object, the quantitative index of neck muscle fatigue, and the head pressure distribution uniformity threshold. Specifically, the signal acquisition method, feature extraction method, and fatigue recognition logic of the biofeedback module in this application mainly acquire neck muscle electrophysiological signals through a brain-computer interface or a high-precision physiological signal monitoring unit. This significantly improves the accuracy and reliability of physiological signal acquisition and avoids signal interference or data distortion problems caused by ordinary monitoring equipment. The biofeedback module actively extracts the temporal and frequency domain features of electromyographic signals. Compared with single feature analysis, it can more comprehensively capture the physiological changes related to muscle fatigue, providing multi-dimensional data support for fatigue state recognition. Based on a preset multi-dimensional threshold system, the fatigue state is quantitatively identified by combining the continuous wearing time threshold, the quantification index of neck muscle fatigue, and the head pressure distribution uniformity threshold, avoiding the one-sidedness of single threshold judgment and making the fatigue state recognition results more scientific and accurate. This application effectively solves the problems of insufficient accuracy and low reliability of fatigue recognition in the prior art, and can provide accurate decision-making basis for subsequent structural parameter adjustment, ensuring that the structural parameter adjustment is highly matched with the actual fatigue state of the target object, thereby more effectively alleviating muscle fatigue and ensuring operational safety and comfort in long-term work scenarios.

[0025] The structural parameters corresponding to the structural parameter adjustment instructions include the ventilation volume, padding hardness, padding fit, head support point position, and wearing tightness of the wearing equipment. In the structural parameter adjustment instructions output by the biofeedback module, the adjustment range of each structural parameter forms a dynamic matching relationship with the quantified fatigue level. Specifically, compared to single or a few parameter adjustments in existing technologies, this application can more systematically improve the wearing experience. The structural parameter adjustment instructions output by the biofeedback module dynamically match the adjustment range of each structural parameter with the quantified fatigue level, achieving personalized control with different adjustment ranges for different fatigue levels. This avoids the limitations of traditional static or fixed-range adjustments and can accurately optimize the fit based on the real-time fatigue state of the target user. For example, for mild fatigue, the padding fit and ventilation volume can be adjusted appropriately; for severe fatigue, the head support point position and wearing tightness can be optimized simultaneously, ensuring the targeted and effective nature of the adjustment measures, significantly improving comfort during long-term wear, and reducing the impact of muscle fatigue accumulation on operational safety.

[0026] Historical usage data includes baseline physiological characteristics of the target object, a dataset of head pressure distribution changes, dynamic curves of neck muscle fatigue, historical records of structural parameter adjustments, environmental temperature and humidity adaptation data, iterative records of personalized ergonomic adaptation parameter sets, and feedback data on ergonomic optimization effects. The data storage module stores all historical usage data in a structured format. Specifically, this application comprehensively accumulates the individual physiological characteristics, usage habits, and adaptation history of the target object, providing rich historical reference for personalized optimization. The data storage module stores all historical usage data in a structured format, ensuring data standardization and reusability, avoiding the problem of ineffective access caused by data chaos, and enabling deep learning algorithms to efficiently mine patterns and correlations in historical data. Through the continuous accumulation and structured storage of historical usage data, the personalized ergonomic adaptation parameter set generated by the optimization decision module can more accurately match the long-term usage habits and physiological characteristic changes of the target object, achieving a better fit with use. At the same time, it provides sufficient data support for the model iteration of deep learning algorithms, continuously improving the optimization accuracy and adaptation performance of the system.

[0027] The personalized ergonomic fitting parameter set includes the size specifications of the wearable equipment, the layout of the ventilation openings, the selection of padding materials, the design of the head support structure, the weight distribution optimization parameters, and the planning of pressure dispersion areas. Specifically, compared with the single-dimensional fitting solutions in existing technologies, this application can achieve a comprehensive and precise fit between the equipment and the target object. All parameters are calibrated according to national ergonomic standards, ensuring the scientific and compliant nature of the fitting solution and avoiding the problem of personalized adjustments being out of ergonomic safety standards. This application enables the personalized ergonomic fitting parameter set to not only meet the individual physiological differences of the target object but also comply with industry-standard safety and comfort. It can provide customized fitting solutions for the high mobility requirements of fighter pilot helmets, the extreme environment adaptation requirements of spacesuit helmets, and the protection and convenience requirements of industrial safety helmets, significantly improving the fit and safety of equipment in different scenarios.

[0028] The cross-modal control module supports multiple operation modes, including voice command recognition, gesture motion capture, and eye tracking. Specifically, these modes break through the limitations of traditional single-channel operations such as buttons and touch controls, providing more flexible and convenient operation options for users. This is particularly suitable for scenarios involving heavy equipment and limited hand movement, such as those experienced by fighter pilots and astronauts. System function commands can be triggered without manual contact with operating components, reducing operational difficulty and response time, and improving safety and convenience. The multi-channel operation design allows users to choose the optimal operation method based on the actual work scenario and their own operating habits. For example, during flight, adjustment commands can be triggered via voice or eye tracking, avoiding distractions from hand operations. In industrial operations, gestures can be used to quickly confirm parameters, significantly improving the system's operational flexibility and practicality, enhancing user experience and operational efficiency.

[0029] Wearable equipment includes, but is not limited to, fighter pilot helmets, spacesuit helmets, or industrial safety helmets. Specifically, for fighter pilot helmets, the system can adapt to the frequent changes in head posture during highly mobile operations, dynamically adjusting support points and pressure distribution to alleviate neck muscle fatigue; for spacesuit helmets, it can cope with extreme environmental temperature and humidity fluctuations, optimizing ventilation and sealing performance while balancing comfort and protection; for industrial safety helmets, it can address the issue of fatigue accumulation during long periods of work, improving wearing comfort and ease of operation through personalized parameter adjustments. The system can quickly adapt to the ergonomic optimization needs of different fields without requiring significant modifications to different equipment types, effectively solving the problems of excessive specialization and narrow applicability in existing technologies, significantly broadening the system's application scenarios, and enhancing the technology's practicality and promotional value.

[0030] According to another aspect of this application, a dynamic ergonomic optimization method based on a dynamic ergonomic optimization system is also provided, comprising the following steps: S1) The system’s multimodal sensing module is activated, which collects head pressure distribution data, neck muscle electrophysiological signals and temperature and humidity parameters of the target object’s environment in real time through integrated sensors, and transmits the collected data synchronously to the biofeedback module, optimization decision module and data storage module. S2) The biofeedback module performs noise filtering, signal amplification and feature normalization preprocessing on the received neck muscle electrophysiological signals and head pressure distribution data. It identifies the fatigue state by combining the continuous use time of the target object, the quantitative index of neck muscle fatigue and the head pressure distribution uniformity threshold, generates the corresponding structural parameter adjustment instructions and sends them to the optimization decision module. S3) The optimization decision module calls a deep learning algorithm pre-trained with historical usage data, integrates real-time environmental temperature and humidity parameters, fatigue status information and historical usage data in the data storage module, and generates a personalized ergonomic adaptation parameter set containing multiple adaptation parameters in accordance with ergonomic standards. S4) The AR feedback module provides real-time feedback on the structural parameter adjustment effect and the adaptation status of the personalized ergonomic adaptation parameter set to the target object in a visual form. S5) The target object triggers a manual adjustment command through the multi-channel operation mode of the cross-modal control module. After receiving the manual adjustment command, the optimization decision module updates the personalized ergonomic adaptation parameter set and simultaneously stores the updated personalized ergonomic adaptation parameter set to the data storage module, forming a dynamic optimization closed loop.

[0031] Specifically, steps S1 to S5 construct a complete dynamic optimization closed loop. Step S1 achieves synchronous acquisition and transmission of multi-source data, ensuring data real-time performance and integrity. Step S2 performs noise filtering, signal amplification, and feature normalization preprocessing on physiological signals, effectively improving signal quality and reducing the impact of interference factors on fatigue recognition results. Step S3 calls a deep learning algorithm pre-trained on historical data to fuse multi-source data and generate a personalized ergonomic adaptation parameter set, ensuring the accuracy and relevance of the adaptation scheme. Step S4 achieves visualized real-time feedback through an AR feedback module, allowing the target object to promptly grasp the adjustment effect and providing a basis for manual intervention. Step S5 supports manual adjustment command triggering and parameter set update storage, forming a cycle from automatic optimization to manual verification and iterative updates. This method has a complete process and rigorous logic, realizing full-process dynamic optimization from data acquisition to scheme optimization and continuous iteration. It avoids the problems of optimization lag and incomplete closed loop in existing technologies, ensuring the timeliness, accuracy, and continuity of ergonomic optimization, and significantly improving the user experience and operational safety of the wearable equipment.

[0032] The biofeedback module can determine fatigue status based on either neck muscle electrophysiological signals or head pressure distribution data, or by combining the two types of data. If based solely on the electrophysiological signals of the neck muscles, the focus is on analyzing the coefficient of variation of signal amplitude and the proportion of spectral energy. If the two types of data are combined, the head pressure distribution uniformity index and the time-domain characteristics of electromyography signals are integrated. The fatigue levels defined by the biofeedback module correspond to preset structural parameter adjustment threshold ranges, and the adjustment thresholds can be dynamically calibrated based on the baseline data of the target object's physiological characteristics. The fatigue state recognition thresholds corresponding to different judgment methods can be calibrated independently.

[0033] Specifically, this application provides a flexible fatigue state assessment method, allowing the biofeedback module to make a judgment based solely on one or a combination of two data sources: neck muscle electrophysiological signals and head pressure distribution data. This adapts to data acquisition conditions in different usage scenarios. For example, in some scenarios, even if a certain type of sensor malfunctions or cannot effectively acquire data, the system can still complete fatigue identification using a single data source, ensuring the system's reliability and adaptability. It clarifies the analytical focus under different assessment methods, making signal analysis more targeted and improving the efficiency and accuracy of fatigue identification. Simultaneously, the fatigue level corresponds to a preset structural parameter adjustment threshold range, and the adjustment threshold can be dynamically calibrated based on the target object's physiological characteristic baseline data. This avoids the limitations of fixed thresholds for different individuals, making structural parameter adjustments more closely aligned with individual differences. This application significantly improves the flexibility, reliability, and accuracy of the biofeedback module, ensuring that fatigue state identification and structural parameter adjustment can adapt to the needs of different scenarios and individuals, further optimizing the system's dynamic adaptability and user experience.

[0034] Example 1: Dynamic Ergonomics Optimization System and Application Method for Fighter Jet Pilot Helmet I. System Configuration Parameters 1. Multimodal sensing module: integrates a high-precision piezoresistive pressure sensor (acquisition accuracy ±0.1kPa, sampling frequency 50Hz), a surface electromyography sensor (signal bandwidth 20-500Hz, common mode rejection ratio ≥80dB), and a digital temperature and humidity sensor (temperature accuracy ±0.3℃, humidity accuracy ±2%RH). All sensors are embedded in the inner lining and outer shell of a fighter pilot's helmet. The pressure sensor has 16 acquisition points covering key stress areas of the head.

[0035] 2. Biofeedback Module: The brain-computer interface is used to collect electrophysiological signals of the trapezius muscle in the neck. In the signal preprocessing stage, a filter is used to filter noise (cutoff frequency 30Hz). The time domain features (root mean square value, integral electromyography value) and frequency domain features (average power frequency, peak frequency) are extracted through feature engineering. A multi-dimensional threshold system is preset: continuous wearing time threshold of 120 minutes, neck muscle fatigue quantification index threshold (root mean square value ≥ 3.5μV), and head pressure distribution uniformity threshold (pressure standard deviation ≤ 2.0kPa).

[0036] 3. Optimize the decision-making module: Convolutional Neural Network (CNN) is selected as the deep learning algorithm. The model input dimension is 256 dimensions (including 16-dimensional stress data, 8-dimensional electromyographic features, 2-dimensional environmental parameters, and 230-dimensional historical data features). The pre-training dataset contains baseline physiological feature data of 100 pilots (average head circumference 56-62cm) and 5000 hours of historical usage data. The initial recognition accuracy of the model is ≥92%.

[0037] 4. AR Feedback Module: Integrated into the built-in miniature AR display of the helmet visor, it displays in real time a heat map of head pressure distribution, muscle fatigue value (0-10 points), comparison curves of structural parameters before and after adjustment, and fit score (0-100 points).

[0038] 5. Cross-modal control module: Supports voice command recognition, gesture motion capture (recognition range 0.3-1.5m, frame rate 30fps), and eye tracking (gaze accuracy ≤0.5°), adaptable to scenarios where pilots wear gloves and operate under high G-forces.

[0039] 6. Data storage module: It adopts distributed database storage, stores data in a structured format, and supports multi-dimensional retrieval of historical data by timestamp, fatigue level, and environmental parameters.

[0040] 7. Personalized Ergonomic Adaptation Parameter Set: Includes helmet size specifications (L size, suitable for head circumference 58-60cm), ventilation vent array layout (6 ventilation vents, 8mm in diameter, distributed on the top and sides), padding material selection (polyurethane foam, density 0.3g / cm³), head support structure design parameters (3 support points, three-dimensional coordinates are X10Y25Z30, X-10Y25Z30, X0Y-30Z25, unit mm), weight distribution optimization coefficient (front side weight ratio ≤45%), and pressure dispersion area planning parameters (forehead and occipital pressure dispersion area ≥15cm²).

[0041] II. Application Method Implementation Steps S1: After the pilot puts on the helmet, the system is activated. The multimodal sensing module simultaneously collects head pressure distribution data (initial average pressure 1.8 kPa), neck muscle electrophysiological signals (initial root mean square value 1.2 μV), and cockpit environmental temperature and humidity parameters (temperature 22℃, humidity 55%RH). The data is transmitted in real time to the biofeedback module, optimization decision module and data storage module.

[0042] S2: The biofeedback module performs noise filtering, signal amplification (1000x amplification) and feature normalization on the electromyographic signal. Combined with the pilot's continuous wearing time (initially 0 minutes), it identifies the fatigue state as no fatigue and generates initial structural parameter adjustment instructions (ventilation volume 5L / min, padding hardness Shore A35, wearing tightness 55N).

[0043] S3: The optimization decision module calls the pre-trained CNN algorithm, integrates real-time environmental parameters, fatigue-free state information and the pilot's historical usage data (pressure distribution records of the previous 30 flights, muscle fatigue change curves) in the data storage module, and generates a personalized ergonomic adaptation parameter set with reference to GB / T 12367-2017 ergonomics standard.

[0044] S4: The AR feedback module visualizes parameters such as ventilation volume and padding hardness on the helmet visor, as well as a pressure distribution heat map (green indicates suitable pressure), with a fit score of 92.

[0045] S5: The pilot triggers a manual adjustment command by saying "increase ventilation" via voice. After receiving the command, the optimization decision module updates the ventilation to 7L / min and stores it in the data storage module. After 2 hours of flight, the multimodal perception module collects the root mean square value of the neck muscle electrophysiological signal, which rises to 3.6μV and the standard deviation of the head pressure distribution is 2.1kPa. The biofeedback module identifies mild fatigue and generates adjustment commands (the padding hardness is reduced to Shore A32, and the head support points are X10Y25Z32, X-10Y25Z32, and X0Y-30Z27). The optimization decision module updates the parameter set after fusing the data, and the AR feedback module displays an adaptation fit score of 95, forming a dynamic optimization closed loop.

[0046] S6: The data storage module continuously accumulates flight data, and the deep learning algorithm is iteratively optimized through offline batch training, improving the fatigue state recognition accuracy to 96% and the average fit of the personalized adaptation parameter set to 3%. III. Implementation Results

[0047] In this embodiment, the system reduces neck muscle fatigue by 40%, improves head pressure distribution uniformity by 35%, and shortens operation response time by 20% after long-term flight (4 hours) by multimodal data fusion and dynamic adjustment. This effectively avoids fatigue accumulation caused by static helmets and ensures operational safety and comfort in high-G, long-term flight scenarios.

[0048] Example 2: Dynamic Ergonomics Optimization System and Application Method for Industrial Safety Helmets I. System Configuration Parameters 1. Multimodal sensing module: integrates a thin-film pressure sensor (acquisition accuracy ±0.2kPa, sampling frequency 30Hz), a flexible electromyography sensor (signal bandwidth 10-400Hz, common mode rejection ratio ≥75dB), and an industrial-grade temperature and humidity sensor (temperature measurement range -10℃~60℃, humidity measurement range 10%RH~90%RH). The sensors are embedded in the lining of an industrial safety helmet, and the pressure sensor has 12 acquisition points.

[0049] 2. Biofeedback Module: Employs a high-precision physiological signal monitoring unit to collect electrophysiological signals of the deltoid muscle in the neck. Fatigue assessment can be performed based solely on electromyography (EMG) signals or in combination with head pressure distribution data. When based solely on EMG signals, the focus is on analyzing the signal amplitude variation coefficient (threshold ≥ 0.3) and the proportion of spectral energy (100-200Hz band proportion ≥ 40%). When combining data, the module integrates the head pressure distribution uniformity index (pressure variation coefficient ≤ 0.25) and the temporal characteristics of EMG signals (integrated EMG value ≥ 50 μV·s), with a multi-dimensional threshold system: a continuous wearing time threshold of 180 minutes and a threshold for quantified neck muscle fatigue indicators (amplitude variation coefficient ≥ 0.35).

[0050] 3. Optimize the decision-making module: Recurrent Neural Network (RNN) is selected as the deep learning algorithm, which supports online incremental training. The model input dimension is 192-dimensional. The pre-training dataset contains baseline physiological characteristic data of 200 industrial workers (average head circumference 55-63cm) and 2000 hours of historical usage data. The initial fatigue recognition accuracy is ≥90%.

[0051] 4. AR Feedback Module: Integrated into a small AR display screen on the front of the helmet, it displays the muscle fatigue level (no fatigue / mild / moderate / severe), current values ​​of structural parameters, ambient temperature and humidity, and fit score in real time.

[0052] 5. Cross-modal control module: Supports voice command recognition (noise immunity ≥85dB) and gesture motion capture (recognizes 5 preset operation gestures), adapts to noisy industrial workshop environments and operation scenarios where hands are wearing gloves, and has a command response time ≤800ms.

[0053] 6. Data storage module: It adopts a hybrid local and cloud storage architecture, with real-time synchronization and updates in the cloud. Data is stored in categories such as "physiological data - environmental data - adjustment records - parameter sets", and the update frequency is once every 1 minute.

[0054] 7. Personalized Ergonomic Adaptation Parameter Set: Includes helmet size specifications (M size, suitable for head circumference 56-58cm), ventilation vent array layout (4 ventilation vents, 10mm in diameter, distributed on the top side), padding material selection (silicone material, Shore A30 hardness), head support structure design parameters (2 support points, 3D coordinates X8Y20Z28, X-8Y20Z28, unit mm), weight distribution optimization coefficient (front side weight ratio ≤42%), and pressure dispersion area planning parameters (temporal and occipital pressure dispersion area ≥12cm²).

[0055] II. Application Method Implementation Steps S1: After industrial workers put on their helmets, they start the system. The multimodal sensing module collects head pressure distribution data (initial average pressure 2.0 kPa), neck muscle electrophysiological signals (initial amplitude variation coefficient 0.15), and workshop environmental temperature and humidity parameters (temperature 28℃, humidity 65%RH) in real time. The data is then transmitted synchronously to each core module.

[0056] S2: The biofeedback module selects and processes head pressure distribution data and electromyography signals. After noise filtering, signal amplification (800x amplification) and feature normalization, combined with the continuous wearing time (initially 0 minutes), it identifies the fatigue state as no fatigue and generates structural parameter adjustment instructions (ventilation volume 6L / min, padding hardness Shore A30, wearing tightness 50N).

[0057] S3: The optimization decision module calls the RNN algorithm, integrates real-time environmental parameters, fatigue-free state information and historical usage data of the operator in the data storage module (adjustment records of the previous 50 operations, fatigue change curves), and generates a personalized ergonomic adaptation parameter set with reference to GB / T 28181-2011 Industrial Helmet Ergonomics Standard.

[0058] S4: The AR feedback module shows that the current fatigue level is no fatigue, the fit score is 88 points, and the values ​​of parameters such as ventilation volume and padding hardness are updated in real time.

[0059] S5: After 3 hours of operation, the multimodal sensing module collected the amplitude variation coefficient of the neck muscle electrophysiological signal, which increased to 0.36, the standard deviation of the head pressure distribution was 2.3 kPa, and the ambient temperature and humidity changed to 30℃ and 68%RH. The biofeedback module identified mild fatigue and generated adjustment instructions (increase ventilation to 8L / min, reduce padding hardness to Shore A28, and tightening to 45N). The optimization decision module updated the parameter set after integrating the data. The operator triggered the manual confirmation instruction by making a gesture (clenching a fist for 3 seconds). The optimization decision module simultaneously stored the updated parameter set to the data storage module.

[0060] S6: During continuous system operation, the deep learning algorithm is iteratively optimized through online incremental training. For every 500 hours of accumulated data, the accuracy of fatigue state recognition increases by 1-2%. By the 100th job, the fit score is stable at over 94 points, forming a complete closed loop from data collection to recognition, decision-making, feedback, and iteration. III. Implementation Results

[0061] In this embodiment, the system adapts to the temperature and humidity fluctuations and long-term working scenarios in industrial workshops, reducing the cumulative amount of neck muscle fatigue by more than 30% after workers wear it continuously for 6 hours, reducing the complaint rate of head discomfort by 50%, and improving the personalization matching degree of the adaptation parameters by more than 40%. It effectively solves the problems of poor adaptability and limited fatigue relief effect of traditional industrial safety helmets, and significantly improves the comfort and safety of industrial operations.

[0062] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A dynamic ergonomic optimization system based on multimodal sensing and biofeedback, characterized in that, include: The multimodal sensing module integrates a pressure sensor, an electromyography (EMG) sensor, and an environmental temperature and humidity sensor. The pressure sensor is suitable for real-time acquisition of head pressure distribution data of the target object, the EMG sensor is suitable for real-time acquisition of neck muscle electrophysiological signals of the target object, and the environmental temperature and humidity sensor is suitable for real-time acquisition of environmental temperature and humidity parameters of the environment in which the target object is located. A biofeedback module is communicatively connected to the multimodal sensing module. The biofeedback module is adapted to receive the electrophysiological signals of the neck muscles and the head pressure distribution data, identify the fatigue state of the target object through physiological signal analysis, and generate structural parameter adjustment instructions. The optimization decision module is communicatively connected to the multimodal perception module and the biofeedback module. The optimization decision module is suitable for calling deep learning algorithms, integrating real-time collected environmental temperature and humidity parameters, fatigue state information and historical usage data, and generating a personalized ergonomic adaptation parameter set in combination with ergonomic standards. The AR feedback module is communicatively connected to the optimization decision module. The AR feedback module is suitable for displaying the effect of structural parameter adjustment and the adaptation status of personalized ergonomic adaptation parameter set in real time. A cross-modal control module is adapted to receive system function commands triggered by a target object through a multi-channel operation mode, and the cross-modal control module is adapted to perform manual interactive control of system parameters and verification and calibration of adaptation parameters; The data storage module is suitable for storing various types of data collected in real time, fatigue state identification results, structural parameter adjustment records, personalized ergonomic adaptation parameter sets, and historical usage data. The data storage module provides data support for the model iterative optimization of deep learning algorithms and the dynamic updating of personalized ergonomic adaptation parameter sets.

2. The dynamic ergonomic optimization system based on multimodal perception and biofeedback according to claim 1, characterized in that, The biofeedback module collects the electrophysiological signals of the neck muscles of the target object through a brain-computer interface or a high-precision physiological signal monitoring unit. The biofeedback module uses feature engineering processing methods to extract the temporal and frequency domain features of the electromyographic signals. The biofeedback module completes the quantitative identification of fatigue state based on a preset multi-dimensional threshold system. The multi-dimensional threshold system includes the continuous wearing time threshold of the target object, the quantitative index of neck muscle fatigue, and the head pressure distribution uniformity threshold.

3. The dynamic ergonomic optimization system based on multimodal perception and biofeedback according to claim 1, characterized in that, The structural parameters corresponding to the structural parameter adjustment instructions include the ventilation volume of the wearable equipment, the hardness of the padding, the fit of the padding, the position of the head support point, and the tightness of the fit. In the structural parameter adjustment instructions output by the biofeedback module, the adjustment range of each structural parameter forms a dynamic matching relationship with the fatigue state quantification level.

4. The dynamic ergonomic optimization system based on multimodal perception and biofeedback according to claim 1, characterized in that, The historical usage data includes baseline data of the target object's physiological characteristics, dataset of changes in head pressure distribution, dynamic curve of neck muscle fatigue, historical records of structural parameter adjustments, environmental temperature and humidity adaptation data, iterative records of personalized ergonomic adaptation parameter sets, and feedback data on ergonomic optimization effects. The data storage module stores all historical usage data in a structured format.

5. The dynamic ergonomic optimization system based on multimodal perception and biofeedback according to claim 1, characterized in that, The personalized ergonomic adaptation parameter set includes the size specifications of the wearable equipment, the layout of the ventilation openings, the selection of the padding material, the design of the head support structure, the weight distribution optimization parameters, and the planning of the pressure dispersion area.

6. The dynamic ergonomic optimization system based on multimodal perception and biofeedback according to claim 1, characterized in that, The cross-modal control module supports multiple operation modes, including voice command recognition, gesture motion capture, and eye-tracking.

7. The dynamic ergonomic optimization system based on multimodal perception and biofeedback according to claim 1, characterized in that, The equipment worn includes, but is not limited to, fighter pilot helmets, spacesuit helmets, or industrial safety helmets.

8. A dynamic ergonomic optimization method based on the system described in any one of claims 1-7, characterized in that, Includes the following steps: S1) The system’s multimodal sensing module is activated, which collects head pressure distribution data, neck muscle electrophysiological signals and temperature and humidity parameters of the target object’s environment in real time through integrated sensors, and transmits the collected data synchronously to the biofeedback module, optimization decision module and data storage module. S2) The biofeedback module performs noise filtering, signal amplification and feature normalization preprocessing on the received neck muscle electrophysiological signals and head pressure distribution data. It identifies the fatigue state by combining the continuous use time of the target object, the quantitative index of neck muscle fatigue and the head pressure distribution uniformity threshold, generates the corresponding structural parameter adjustment instructions and sends them to the optimization decision module. S3) The optimization decision module calls a deep learning algorithm pre-trained with historical usage data, integrates real-time environmental temperature and humidity parameters, fatigue status information and historical usage data in the data storage module, and generates a personalized ergonomic adaptation parameter set containing multiple adaptation parameters in accordance with ergonomic standards. S4) The AR feedback module provides real-time feedback on the structural parameter adjustment effect and the adaptation status of the personalized ergonomic adaptation parameter set to the target object in a visual form. S5) The target object triggers a manual adjustment command through the multi-channel operation mode of the cross-modal control module. After receiving the manual adjustment command, the optimization decision module updates the personalized ergonomic adaptation parameter set and simultaneously stores the updated personalized ergonomic adaptation parameter set to the data storage module, forming a dynamic optimization closed loop.

9. The dynamic ergonomic optimization method based on multimodal perception and biofeedback according to claim 8, characterized in that, The biofeedback module can determine fatigue status based on either neck muscle electrophysiological signals or head pressure distribution data, or by combining the two types of data. If based solely on the electrophysiological signals of the neck muscles, the focus is on analyzing the coefficient of variation of signal amplitude and the proportion of spectral energy. If the two types of data are combined, the head pressure distribution uniformity index and the time-domain characteristics of electromyography signals are integrated. The fatigue levels defined by the biofeedback module correspond to preset structural parameter adjustment threshold ranges, and the adjustment thresholds can be dynamically calibrated based on the baseline data of the target object's physiological characteristics. The fatigue state identification thresholds corresponding to different judgment methods can be calibrated independently.

Images