Multi-modal physiological sensing based assistive and cooling collaborative control method and system

Abstract

The application relates to the technical field of intelligent construction, and specifically discloses a power-assisted and cooling collaborative control method and system based on multi-modal physiological sensing, which specifically comprises the following steps: synchronously collecting electrocardiosignal or photoplethysmographic pulse wave signal, surface electromyogram, IMU posture, skin temperature and environmental temperature and humidity signals of a worker; performing a taking operation on the collected signals to output a standardized feature vector; calculating a comprehensive fatigue index F based on the output standardized feature vector; calling an individual fatigue baseline established according to historical operation data of the worker to solve a real-time fatigue deviation value; synchronously calculating a power-assisted gain coefficient and a cooling gear based on the real-time fatigue deviation value; and dynamically adjusting the joint output torque of an exoskeleton and the refrigeration output power according to the power-assisted gain coefficient and the cooling gear respectively. The application can synchronously link and adjust the power-assisted output and the cooling gear according to the real-time fatigue of the human body, so that power assistance and cooling are collaboratively controlled.

Description

Technical Field

[0001] This invention relates to the field of intelligent construction technology, and more specifically, to a method and system for coordinated control of assistance and cooling based on multimodal physiological perception. Background Technology

[0002] In construction, logistics, and heavy manufacturing settings, workers are exposed to high temperatures and heavy workloads for extended periods, facing multiple safety risks including musculoskeletal injuries, cardiovascular heat stress, and heatstroke. Existing protective equipment falls into two categories, both with significant technical shortcomings: 1. Active-assisted exoskeleton: Relying solely on joint torque and IMU motion signals to output assist torque in an open loop, it completely fails to incorporate feedback from the human body's physiological state. It cannot distinguish between a worker's "light workload with low fatigue" and "continuous heavy load with extreme fatigue," and cannot automatically increase assist to reduce load when fatigue accumulates; continuous high assist under low load results in energy waste, and it does not consider the cardiovascular stress load caused by high temperatures.

[0003] 2. Wearable cooling equipment (phase change / liquid cooling vest): The cooling power is manually / automatically adjusted based on the ambient temperature and humidity. A single environmental parameter cannot characterize the individual's physiological heat tolerance differences. It is worn separately from the exoskeleton. The two power supplies and two control logics are not linked to each other, resulting in structural defects such as the external vest interfering with joint movement, excessive overall weight, and short battery life of the two sets of batteries.

[0004] 3. Limitations of existing physiological monitoring exoskeletons: They only collect single electromyographic or heart rate signals, lack HRV heart rate variability to characterize autonomic / cardiovascular load, and cannot distinguish between simple exercise fatigue and heat stress-induced cardiovascular overload; there is no personalized baseline correction, and the control parameters are uniform for different physical workers under the same load, resulting in poor adaptability; a unified fatigue index has not been established to simultaneously drive two sets of actuators for assistance and cooling, and can only intervene in muscle load or thermal load individually.

[0005] In summary, existing technologies lack multimodal four-dimensional physiological fusion assessment, synchronous and coordinated regulation of assist and cooling, and integrated solutions for structure and power supply, and cannot meet the comprehensive fatigue protection needs of construction workers in high-temperature and heavy-load environments. Summary of the Invention

[0006] The technical problem to be solved by this invention is to provide a method and system for coordinated control of assistance and cooling based on multimodal physiological perception; The solution adopted by this invention to solve the technical problem is: A multimodal physiological sensing-based assist and cooling coordinated control system, used in conjunction with an exoskeleton, includes a data acquisition module for collecting workers' electrocardiogram or photoplethysmography (PPG) signals, surface electromyography (EMG) signals, posture signals, body temperature, and ambient temperature and humidity; a signal preprocessing and feature extraction module connected to the data acquisition module for filtering, denoising, feature extraction, and normalization of the received data and outputting standardized feature vectors; a multimodal fatigue index calculation module connected to the signal preprocessing and feature extraction module for calculating a comprehensive fatigue index; a personalized fatigue baseline matching module connected to the multimodal fatigue index calculation module; a coordinated control decision module connected to the personalized fatigue baseline matching module; an assist drive module and a cooling control module connected to the coordinated control decision module; and a cooling module connected to the cooling control module.

[0007] In some possible implementations, the data acquisition module includes a heart rate variability sensing module, a surface electromyography sensing module, an IMU posture sensing module, and a temperature sensing module, which are respectively connected to the signal preprocessing and feature extraction modules.

[0008] In some possible implementations, an energy management module is also included for supplying power and managing the power distribution of the data acquisition module, signal preprocessing and feature extraction module, multimodal fatigue index calculation module, personalized fatigue baseline matching module, assist drive module, collaborative control decision module and cooling control module.

[0009] on the other hand: This invention also provides a method for coordinated control of assistance and cooling based on multimodal physiological perception. The coordinated control system for assistance and cooling described above specifically includes the following steps: Step S1: Synchronously acquire the worker's electrocardiogram or photoplethysmography pulse wave signal, surface electromyography, IMU posture, skin temperature and ambient temperature and humidity signals; Step S2: Filter, denoise, time-series alignment and feature extraction are performed on the collected signal, and the extracted physiological features are uniformly mapped to the [0,1] interval using the extreme value normalization method, and a standardized feature vector is output. Step S3: Based on the output standardized feature vector, calculate the comprehensive fatigue index F through a four-component weighted fusion model, where the value of F ranges from [0, 1], where 0 represents a fully conscious state and 1 represents a state of extreme fatigue. Step S4: Retrieve the individual fatigue baseline established based on the worker's historical work data, and solve for the real-time fatigue deviation value; Step S5: Calculate the booster gain coefficient and cooling level synchronously based on the real-time fatigue deviation value; Step S6: Dynamically adjust the joint output torque and cooling output power of the exoskeleton according to the assist gain coefficient and cooling level; Step S7: Repeat steps S1 to S6, and simultaneously reduce the power of assist and cooling when the fatigue index drops.

[0010] In some possible implementations, in step S3, calculating the comprehensive fatigue index F specifically refers to: (1); in, Cardiovascular component weighting ； This is a component related to cardiovascular fatigue. For muscle component weights, ; This is the muscle fatigue component; For attitude component weights, ; This refers to the attitude load component; For the thermal stress component weighting, ; This is the thermal stress component; 。

[0011] In some possible implementations, step S4 specifically refers to, (2); in, This represents the real-time fatigue deviation value. As the baseline for individual fatigue, .

[0012] In some possible implementations, the boost gain coefficient in step S5 is calculated as follows: (3); in, To boost the gain coefficient; Basic boost; To assist in gain adjustment, .

[0013] In some possible implementations, the calculation of the cooling level in step S5 is as follows: (4); in, This is the rounding function; For the cooling setting, its value is... ； This refers to the cooling level adjustment rate.

[0014] In some possible implementations, during use, the power output is allocated by the energy management module according to the current assist power and cooling power requirements, using the following formula; (5); (6); in, Total battery output power; To boost power; For the target cooling capacity, This refers to the actual cooling power output. This is the maximum output power; when At the time, ensure ,at this time .

[0015] In some possible implementations, in step S6, when adjusting the joint output torque of the exoskeleton, the assist drive module scales the basic assist curve of the exoskeleton according to the assist gain coefficient to obtain the actual assist torque of each joint, and controls the upper limb assist mechanism and / or lower limb assist mechanism of the exoskeleton by outputting the adjusted torque through the drive motor.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention integrates multimodal information such as heart rate variability, surface electromyography, posture, and temperature to calculate a comprehensive fatigue index. It combines this with a personalized fatigue baseline specifically calibrated for each worker to obtain fatigue deviation, thus achieving individualized fatigue assessment rather than a uniform fixed threshold, and adapting to the physiological differences of different workers. By collecting attitude signals to retrieve the pre-stored basic assist torque curve, and by relying on fatigue deviation to solve the assist gain coefficient to scale the existing basic assist curve, the assist output and cooling level are adjusted synchronously according to the real-time fatigue of the human body to achieve coordinated control of assist and cooling.

[0017] By setting up an energy management module and introducing a power constraint formula for priority assistance, in scenarios where the total battery power is limited, priority is given to ensuring the output of the exoskeleton's assistance, while dynamically limiting the maximum output power of cooling, thus avoiding exceeding the power limit of the entire machine and ensuring the safety of heavy-duty operations. Detailed Implementation

[0018] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. The terms "first," "second," and similar terms used in this application do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Similarly, "a" or "one," etc., do not indicate a quantity limitation, but rather indicate the existence of at least one. In the implementation of this application, "and / or" describes the association relationship of related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. In the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more. For example, multiple positioning posts refer to two or more positioning posts. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0019] The present invention will now be described in detail.

[0020] A multimodal physiological sensing-based assist and cooling coordinated control system, used in conjunction with an exoskeleton, includes a data acquisition module for collecting workers' electrocardiogram or photoplethysmography (PPG) signals, surface electromyography (EMG) signals, posture signals, body temperature, and ambient temperature and humidity; a signal preprocessing and feature extraction module connected to the data acquisition module for filtering, denoising, feature extraction, and normalization of the received data and outputting standardized feature vectors; a multimodal fatigue index calculation module connected to the signal preprocessing and feature extraction module for calculating a comprehensive fatigue index; a personalized fatigue baseline matching module connected to the multimodal fatigue index calculation module; a coordinated control decision module connected to the personalized fatigue baseline matching module; an assist drive module and a cooling control module connected to the coordinated control decision module; and a cooling module connected to the cooling control module.

[0021] In some possible implementations, the data acquisition module includes a heart rate variability sensing module, a surface electromyography sensing module, an IMU posture sensing module, and a temperature sensing module, which are respectively connected to the signal preprocessing and feature extraction modules.

[0022] In some possible implementations, an energy management module is also included for supplying power and managing the power distribution of the data acquisition module, signal preprocessing and feature extraction module, multimodal fatigue index calculation module, personalized fatigue baseline matching module, assist drive module, collaborative control decision module and cooling control module.

[0023] Specifically, the heart rate variability (HRV) sensing module is used to collect workers' electrocardiogram (ECG) signals or photoplethysmography (PPG) pulse wave signals, extracting heart rate and HRV indices to assess cardiovascular load and autonomic nervous system regulation. The time-domain and frequency-domain indices of HRV reflect the sympathetic-parasympathetic balance. When continuous work leads to fatigue accumulation, the time-domain indices decrease while the frequency-domain indices increase, indicating increased physiological load. The output heart rate and HRV indices from the HRV sensing module are sent to the signal preprocessing and feature extraction module as input for the cardiovascular dimension of fatigue index calculation.

[0024] The surface electromyography (EMG) sensing module is used to collect surface EMG signals from the shoulder, back, and thigh muscles of workers, extract the median frequency and root mean square (RMS) values ​​of the EMG signals, and use them to assess the degree of local muscle fatigue. When muscles are fatigued, the median frequency of the EMG signals drifts to lower frequencies and the RMS value increases, which can quantitatively characterize the depth of muscle fatigue. The surface EMG sensing module outputs surface EMG signals to the signal preprocessing and feature extraction module, which serves as the muscle dimension input for fatigue index calculation.

[0025] The IMU posture sensing module is used to collect posture angle and angular velocity information of the worker's torso and limbs, and to identify posture signals such as bending over, squatting, raising arms, and maintaining a fixed posture for a long time, as well as the duration of these actions, to assess posture load. Preferably, maintaining an undesirable posture for a long time is considered a high posture load. This module outputs posture features to the signal preprocessing and feature extraction module as the posture dimension input for fatigue index calculation. At the same time, it provides motion intention information to the assist drive module. Maintaining an undesirable posture for a long time, such as bending over >60° for more than 10 minutes, is described here.

[0026] The temperature sensing module collects workers' skin temperature and ambient temperature and humidity to calculate heat stress indicators. Specifically, when the ambient temperature exceeds 35°C and skin temperature continues to rise, it is determined to be a state of high heat stress. The temperature sensing module outputs heat stress indicators to the signal preprocessing and feature extraction module as input for the heat stress dimension of fatigue index calculation; it also provides environmental feedback to the cooling control module.

[0027] The signal preprocessing and feature extraction module performs filtering, noise reduction, feature extraction, and normalization on the above four types of signals, and outputs a standardized feature vector; the heart rate signal (ECG signal or photoplethysmography pulse wave signal) is bandpass filtered and then R-wave detection is performed to calculate the heart rate variability index. The median frequency and root mean square value of surface electromyography (EMG) signals were calculated after bandpass filtering. The attitude signal is filtered by complementary filtering to extract the attitude angle and angular velocity. The temperature signals (body temperature and ambient temperature and humidity) are averaged and then output as steady-state values. The signal preprocessing and feature extraction module receives signals from the heart rate variability sensing module, surface electromyography sensing module, IMU posture sensing module, and temperature sensing module, and outputs standardized feature vectors to the fatigue index calculation module.

[0028] The multimodal fatigue index calculation module calculates the comprehensive fatigue index F based on four types of standardized feature vectors using a weighted fusion model. The multimodal fatigue index calculation module receives the standardized feature vectors output by the signal preprocessing and feature extraction module and outputs the fatigue index to the personalized fatigue baseline matching module.

[0029] The personalized fatigue baseline matching module establishes an individual fatigue baseline model based on workers' historical work data, compares the real-time fatigue index with the individual baseline, and outputs a personalized fatigue deviation value. Since there are significant differences in the baseline heart rate variability and electromyographic response of different workers under the same external load, personalized baseline matching can eliminate the influence of individual differences. It receives the output from the fatigue index calculation module, combines it with the stored individual baseline data, and outputs the personalized fatigue deviation to the collaborative control decision module.

[0030] The collaborative control decision module calculates both the assist gain coefficient and the cooling level command based on personalized fatigue deviations, achieving dual-path collaborative intervention. When the fatigue index increases, both the assist gain and the cooling level increase simultaneously; when the fatigue index decreases, both decrease simultaneously to save energy. The collaborative control decision module receives personalized fatigue deviations, outputs the assist gain coefficient to the assist drive module, and outputs the cooling level command to the cooling control module.

[0031] The assist drive module adjusts the motor output torque according to the assist gain coefficient to drive the upper and / or lower limb assist mechanisms of the exoskeleton; the assist gain coefficient scales and adjusts the pre-stored basic assist curve of the exoskeleton; the assist drive module receives the assist gain coefficient from the collaborative control decision module and the information from the posture sensing module to drive the motor actuators on the back frame of the exoskeleton.

[0032] The cooling control module receives cooling level commands from the collaborative control decision module and controls the cooling and heat dissipation components of the phase change cooling vest. The cooling control module controls the cooling module's operation according to the cooling level commands. The cooling module includes a semiconductor cooling chip and a fan integrated into the phase change cooling vest on the exoskeleton, thereby achieving cooling through PCB cold storage, the semiconductor cooling chip, and the fan.

[0033] During use, the energy management module dynamically allocates battery output according to priority strategies based on the current power assist and cooling power requirements. When the total power demand exceeds the battery output limit, priority is given to ensuring power assist, while cooling is downgraded to PCM passive mode. The system reassesses the fatigue index at a set interval (e.g., 10 seconds). If the fatigue level decreases, the power assist and cooling levels are reduced to save energy; if the fatigue level continues to rise, the current strategy is maintained or enhanced, forming a closed loop.

[0034] on the other hand: This invention also provides a method for coordinated control of assistance and cooling based on multimodal physiological perception. The coordinated control system for assistance and cooling described above specifically includes the following steps: Step S1: Synchronously acquire the worker's electrocardiogram or photoplethysmography pulse wave signal, surface electromyography, IMU posture, skin temperature and ambient temperature and humidity signals; Step S2: Filter, denoise, time-series alignment and feature extraction are performed on the collected signal, and the extracted physiological features are uniformly mapped to the [0,1] interval using the extreme value normalization method, and a standardized feature vector is output. Step S3: Based on the output standardized feature vector, calculate the comprehensive fatigue index F using a four-component weighted fusion model, where F ranges from [0, 1], where 0 represents a fully conscious state and 1 represents a state of extreme fatigue. Specifically, calculating the comprehensive fatigue index F involves: (1); in, Cardiovascular component weighting Preferably, the value is 0.2-0.3. ； This is a component related to cardiovascular fatigue. For muscle component weights, Preferably, the value is 0.3-0.4; This is the muscle fatigue component; For attitude component weights, Preferably, the value is 0.15-0.25; This refers to the attitude load component; For the thermal stress component weighting, Preferably, the value is 0.15-0.25; This is the thermal stress component; 。

[0035] Each component is obtained by normalizing and mapping the corresponding sensing signal: Mapped by the rate of decline of the time-domain indicator, Mapped by the low-frequency drift rate of the median electromyography frequency, Mapped by the duration of poor posture, It is a joint mapping of skin temperature rise and ambient temperature.

[0036] Step S4: Retrieve the individual fatigue baseline established based on the worker's historical work data, and calculate the real-time fatigue deviation value; specifically, this means... (2); in, This represents the real-time fatigue deviation value. As the baseline for individual fatigue, .

[0037] Step S5: Calculate the booster gain coefficient and cooling level synchronously based on the real-time fatigue deviation value; The boost gain coefficient is calculated as follows: (3); in, To assist the gain coefficient, 0.5 ≤ ≤2, assist when there is no fatigue deviation =1; Basic boost; To assist in gain adjustment, Preferably, The range is 1-3.

[0038] The calculation for the cooling level is as follows: (4); in, This is the rounding function; For the cooling setting, its value is... ; in When =0, it relies on PCM passive cold storage for cooling; When =1, it consists of primary low-end semiconductor refrigeration and PCM passive cold storage refrigeration; When =2, it includes mid-range semiconductor cooling, low-temperature air cooling, and PCM passive cold storage cooling; When =3, high-end semiconductor refrigeration, forced air cooling, and PCM passive cold storage refrigeration are used; The preferred value for the cooling level adjustment rate is 3-6.

[0039] Step S6: Dynamically adjust the joint output torque and cooling output power of the exoskeleton according to the assist gain coefficient and cooling level; when adjusting the joint output torque of the exoskeleton, the assist drive module scales the basic assist curve of the exoskeleton according to the assist gain coefficient to obtain the actual assist torque of each joint, and controls the upper limb assist mechanism and / or lower limb assist mechanism of the exoskeleton by outputting the adjusted torque through the drive motor.

[0040] Step S7: Repeat steps S1 to S6, and simultaneously reduce the power of assist and cooling when the fatigue index drops.

[0041] In some possible implementations, during use, the power output is allocated by the energy management module according to the current assist power and cooling power requirements, using the following formula; (5); (6); in, Total battery output power; To boost power; For the target cooling capacity, This refers to the actual cooling power output. This is the maximum output power; when At the time, ensure ,at this time .

[0042] Example 1 This system is adapted to steel structure installation operations on construction sites in summer with temperatures up to 38°C. Targeting typical work conditions such as continuous arm-raising welding and high-altitude bolt tightening, the energy management module has a maximum output power of 400W. The exoskeleton is an upper limb-assisted exoskeleton, specifically using a hybrid drive of passive springs and active motors. The phase change cooling vest has a PCM phase change temperature of 28°C, a semiconductor cooling chip with a rated power of 80W, and a fan with a rated power of 20W. I. Initial Stage (0-30 minutes) Before starting the operation, the worker completed the system quick wear calibration. The heart rate chest strap, which serves as the heart rate variability sensing module, was fixed at the lower 1 / 3 of the sternum. The surface electromyography (SEM) patch, which serves as the surface electromyography (SEM) sensing module, was fixed on the shoulders, waist, and thighs, avoiding the muscle and tendon attachment areas. The torso IMU, which serves as the posture sensing module, was fitted to the center of the semi-rigid exoskeleton back plate. The skin temperature sensor was placed close to the area below the scapula where there are no dense sweat glands. The ambient temperature and humidity sensor was placed in the ventilation gap on the side of the back frame to avoid direct sunlight affecting the sampling accuracy.

[0043] After the operation started, the system acquired electromyography signals at 1000Hz, IMU posture data at 100Hz, and heart rate and temperature / humidity data at 1Hz. After 50Hz notch filtering and bandpass denoising preprocessing, the overall fatigue index F=0.15 was calculated in the initial state, and the individual fatigue baseline was determined. This only represents the basic physiological fluctuations experienced by workers upon entering a high-temperature environment and undergoing slight thermal adaptation. The corresponding individualized fatigue deviation ΔF = 0.03 was calculated, yielding the assist gain coefficient. =1.06, cooling setting =0, the system automatically enters PCM passive cooling mode.

[0044] II. Mid-term stage (30-120 minutes) After continuous arm-raising welding operations, the data acquisition module captured clear physiological and postural change signals: the median frequency (MDF) of the deltoid muscle in the shoulder decreased by 20% compared to the initial state, reflecting significant metabolic fatigue accumulation in the muscles; the RMSSD index of heart rate variability (HRV) decreased by 15%, indicating that the autonomic nervous system began to shift towards sympathetic excitation, and the cardiovascular load gradually increased; the trunk skin temperature increased by 1.2℃ compared to the initial value, and the ambient temperature fluctuated slightly from 38℃ to 38.7℃, indicating that the risk of heat stress began to appear; at the same time, the trunk IMU statistics showed that the cumulative duration of the worker's continuous arm-raising posture reached 22 minutes, and the static muscle maintenance load increased significantly.

[0045] Substituting the four signal components into a preset weighted fusion formula, with a weight of 0.25 for the cardiovascular component, 0.35 for the muscle component, 0.2 for the posture component, and 0.2 for the heat stress component, the comprehensive fatigue index is calculated. F =0.43, corresponding to personalized fatigue deviation ΔF =0.31.

[0046] The system automatically triggers a collaborative control strategy to boost the gain coefficient. =1.62, an improvement of over 60% compared to when there is no fatigue deviation; cooling mode. =1, the system uses low-end semiconductor cooling chips and PCM cooling.

[0047] III. High-load phase (120-180 minutes) After two hours of continuous work, the cumulative effect of fatigue intensified: the RMS amplitude of shoulder electromyography increased by 42% compared to the initial state, and muscle tremor signals began to appear frequently; the heart rate remained above 115 beats / minute, and the HRV fluctuation range further narrowed; the cumulative time for workers to continuously bend over and tighten bolts exceeded 40 minutes, and the lumbar and back load under the forward-leaning posture was significantly increased; the ambient temperature climbed to 39.2℃, and the local surface radiation temperature of the steel structure exceeded 45℃, resulting in a significant increase in the additional heat load from thermal radiation.

[0048] The multimodal sensing module fuses all signal components in real time, calculates the comprehensive fatigue index F=0.72, and the individual fatigue deviation ΔF=0.6. Substituting these values ​​into the control formula, the boost gain coefficient is then calculated. =2.0, reaching the system's preset boost gain limit; cooling level. =2, the system automatically switches to high-end cooling mode: medium-range cooling of semiconductor cooling chip and air cooling start-up auxiliary heat dissipation and cooling.

[0049] At this point, the power assist output reaches its maximum value, the semiconductor cooling is upgraded to medium speed, and the fan starts to assist in heat dissipation. At this time, the total power consumption of the motor on the assist side is 180W, the semiconductor cooling chip on the cooling side operates at 60% of its rated power, and a single 20W air-cooled fan starts at 30% speed. The instantaneous total power consumption of the whole machine is about 234W, and the battery has sufficient remaining power.

[0050] IV. Extreme Phase (180-240 minutes) The ambient temperature rose to 40℃, the surface radiant temperature of the steel structure reached 47℃, the worker's overall fatigue index F=0.91, the individual fatigue deviation ΔF=0.79, the heart rate rose to 132 beats / minute, and the skin temperature exceeded 37.5℃. The assist gain was maintained at its maximum of 2.0, and the cooling setting was... =3, in high-end cooling mode, the semiconductor cooling chip runs at 100% rated power of 80W, the two 20W air-cooled fans run at full speed, the instantaneous total power consumption of the whole machine is 390W, close to the battery output threshold of 400W; the system issues a high fatigue warning and recommends that the worker rest.

[0051] This invention is not limited to the specific embodiments described above. The invention extends to any new feature or combination disclosed in this specification, as well as any new method or process step or combination disclosed herein.

Claims

1. A multimodal physiological sensing-based assistive and cooling coordinated control system, used in conjunction with an exoskeleton, characterized in that, It includes a data acquisition module for collecting workers' electrocardiogram or photoplethysmography (PPG) pulse wave signals, surface electromyography (EMG) signals, posture signals, body temperature, and ambient temperature and humidity; a signal preprocessing and feature extraction module connected to the data acquisition module for filtering, denoising, feature extraction, and normalization of the received data and outputting standardized feature vectors; a multimodal fatigue index calculation module connected to the signal preprocessing and feature extraction module for calculating the comprehensive fatigue index; a personalized fatigue baseline matching module connected to the multimodal fatigue index calculation module; a collaborative control decision module connected to the personalized fatigue baseline matching module; a power assist drive module and a cooling control module connected to the collaborative control decision module respectively; and a cooling module connected to the cooling control module.

2. The assist and cooling coordinated control system based on multimodal physiological perception according to claim 1, characterized in that, The data acquisition module includes a heart rate variability sensing module, a surface electromyography sensing module, an IMU posture sensing module, and a temperature sensing module, which are respectively connected to the signal preprocessing and feature extraction modules.

3. The assist and cooling coordinated control system based on multimodal physiological perception according to claim 1, characterized in that, It also includes an energy management module for powering and managing the power distribution of the data acquisition module, signal preprocessing and feature extraction module, multimodal fatigue index calculation module, personalized fatigue baseline matching module, power assist drive module, collaborative control decision module and cooling control module.

4. A method for coordinated control of assist and cooling based on multimodal physiological perception, characterized in that, The assist and cooling coordinated control system according to any one of claims 1-3 specifically includes the following steps: Step S1: Synchronously acquire the worker's electrocardiogram or photoplethysmography pulse wave signal, surface electromyography, IMU posture, skin temperature and ambient temperature and humidity signals; Step S2: Filter, denoise, time-series alignment and feature extraction are performed on the collected signal, and the extracted physiological features are uniformly mapped to the [0,1] interval using the extreme value normalization method, and a standardized feature vector is output. Step S3: Based on the output standardized feature vector, calculate the comprehensive fatigue index F through a four-component weighted fusion model, where the value of F ranges from [0, 1], where 0 represents a fully conscious state and 1 represents a state of extreme fatigue. Step S4: Retrieve the individual fatigue baseline established based on the worker's historical work data, and solve for the real-time fatigue deviation value; Step S5: Calculate the booster gain coefficient and cooling level synchronously based on the real-time fatigue deviation value; Step S6: Dynamically adjust the joint output torque and cooling output power of the exoskeleton according to the assist gain coefficient and cooling level; Step S7: Repeat steps S1 to S6, and simultaneously reduce the power of assist and cooling when the fatigue index drops.

5. The method for coordinated control of assistance and cooling based on multimodal physiological perception according to claim 4, characterized in that, In step S3, calculating the comprehensive fatigue index F specifically refers to: （1）； in, Cardiovascular component weighting ； This is a component related to cardiovascular fatigue. For muscle component weights, ; This is the muscle fatigue component; For attitude component weights, ; This refers to the attitude load component; For the thermal stress component weighting, ; This is the thermal stress component; 。 6. The method for coordinated control of assistance and cooling based on multimodal physiological perception according to claim 4, characterized in that, Specifically, step S4 refers to... （2）； in, This represents the real-time fatigue deviation value. As the baseline for individual fatigue, .

7. The method for coordinated control of assistance and cooling based on multimodal physiological perception according to claim 6, characterized in that, The calculation of the boost gain coefficient in step S5 is as follows: （3）； in, To boost the gain coefficient; Basic boost; To assist in gain adjustment, .

8. The method for coordinated control of assistance and cooling based on multimodal physiological perception according to claim 6, characterized in that, The calculation of the cooling level in step S5 is as follows: （4）； in, This is the rounding function; For the cooling setting, its value is... ； This refers to the cooling level adjustment rate.

9. The method for coordinated control of assistance and cooling based on multimodal physiological perception according to claim 4, characterized in that, When in use, the power output is allocated according to the current power and cooling power requirements by the energy management module, using the following formula; （5）； （6）； in, Total battery output power; To boost power; For the target cooling capacity, This refers to the actual cooling power output. This is the maximum output power; when At the time, ensure ,at this time .

10. The method for coordinated control of assistance and cooling based on multimodal physiological perception according to claim 3, characterized in that, In step S6, when adjusting the joint output torque of the exoskeleton, the assist drive module scales the basic assist curve of the exoskeleton according to the assist gain coefficient to obtain the actual assist torque of each joint. The adjusted torque is then output through the drive motor to control the upper limb assist mechanism and / or lower limb assist mechanism of the exoskeleton.

Images