Man-machine exoskeleton cooperation method based on multi-modal signal fusion

Abstract

The invention discloses a man-machine exoskeleton cooperation method based on multi-modal signal fusion, and relates to the technical field of exoskeleton program control. The method comprises the following steps that man-machine exoskeleton driving pre-pressure cold and hot mutation analysis is conducted in combination with man-machine exoskeleton man-machine cooperation components; and performing memory alloy spring pre-pressure optimization adjustment in combination with a man-machine exoskeleton driving pre-pressure cold and hot mutation analysis result. According to the invention, cold chain exoskeleton cooperative control is realized through multi-mode signal fusion, and three-level regulation and control are implemented based on a current threshold value and a myoelectricity state: graded current rising / falling and safety early warning are carried out in case of abnormal current; dynamic gain compensation or joint retardation protection is carried out when myoelectricity is abnormal; according to the man-machine exoskeleton cooperative control method and the man-machine exoskeleton cooperative control system, the effect of improving the accuracy of man-machine exoskeleton cooperative control for coping with cold-chain logistics is achieved, and the problem that the accuracy of man-machine exoskeleton cooperative control for coping with cold-chain logistics is disabled in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of exoskeleton program control, and in particular to a human-machine exoskeleton collaboration method based on multimodal signal fusion. Background Art

[0002] Current exoskeleton technologies for environmental perception, motion intention recognition, and powertrain development and validation are largely based on relatively mild and stable laboratory environments. Mainstream solutions rely on combining visual and depth information with an inertial measurement unit (IMU) to achieve simultaneous positioning, mapping, and dynamic obstacle avoidance. Surface electromyography sensors attached to the skin collect muscle electrical signals, decode the user's intended force, and trigger corresponding assistance. Assistance is provided by hydraulic or electric drive systems, with lithium-ion batteries providing energy. Control systems typically operate based on pre-set human motion models and relatively static signal processing algorithms. However, these devices operate under constant, moderate temperature conditions, and the user's physiological state and environment are relatively static. This presents a fundamental mismatch for cold chain handling scenarios, which require frequent movement between -25°C cold storage and normal-temperature loading and unloading areas.

[0003] For example, the invention patent with announcement number: CN117901071B discloses an upper limb exoskeleton robot with active and passive collaborative power assistance, including: in the upper limb structure, a parallelogram mechanism is used to actively match the shoulder rotation point, and the structural characteristics of the parallelogram mechanism itself are combined to achieve a lightweight design, effectively reducing the interference of the exoskeleton on the human body and the load that the human shoulder needs to bear; the upper arm part adopts an active power assistance method of motor-driven flexible cable transmission to help operators achieve on-demand active power assistance according to the power assistance process.

[0004] For example, the invention patent with announcement number: CN118559685B discloses a lower limb exoskeleton robot, including: the lower limb exoskeleton robot is composed of a waist adjustment module, a thigh adjustment module, a calf adjustment module, an ankle joint height adjustment module, a foot adjustment module, a power device and a fixing device. The waist adjustment module and the thigh adjustment module are connected through a power device, the thigh adjustment module and the calf adjustment module are connected through a power device, and the lower end of the calf adjustment module is connected to the foot adjustment module through the ankle joint height adjustment module and the gas spring assist device. Compared with the existing exoskeleton robot, the present invention

[0005] However, in the process of implementing the technical solutions of the invention in the embodiments of the present application, the present application found that the above technology has at least the following technical problems:

[0006] In the existing technology, the existing cold chain logistics exoskeleton technology faces the comprehensive problems of mutual coupling and vicious cycle of multiple technical links under extreme conditions in actual application. In order to prevent condensation on the sensor lens, the battery life will plummet; attempts to increase the battery life will make the equipment too heavy; workers' sweat seriously interferes with the already difficult biological signal collection, and intention recognition failure leads to insufficient assistance; materials repeatedly expand and contract in frequent and drastic temperature jumps; low temperatures cause workers' muscles to stiffen, and the electromyographic signals are weak and the characteristics change, but the machine algorithm still uses the normal temperature model; the cross-coupling of comprehensive factors has jointly created a systemic problem of control disability, and there is a problem of human-machine exoskeleton collaborative control accuracy disability in response to cold chain logistics. Summary of the Invention

[0007] The embodiment of the present application solves the problem in the prior art of the inability of human-machine-exoskeleton collaborative control to cope with cold chain logistics by providing a human-machine-exoskeleton collaborative method based on multimodal signal fusion, thereby achieving the effect of improving the accuracy of human-machine-exoskeleton collaborative control to cope with cold chain logistics.

[0008] An embodiment of the present application provides a human-machine-exoskeleton collaboration method based on multimodal signal fusion, comprising the following steps: obtaining a human-machine-exoskeleton human-machine collaboration component through human-machine collaboration data coupling analysis; performing a cold and hot mutation analysis of a human-machine-exoskeleton drive pre-stress in combination with the human-machine-exoskeleton human-machine collaboration component; performing electromyographic mutation analysis to optimize and adjust the human-machine-exoskeleton human-machine collaboration component or issue an early warning; and optimizing and adjusting the pre-stress of a memory alloy spring in combination with the results of the cold and hot mutation analysis of a human-machine-exoskeleton drive pre-stress.

[0009] Furthermore, a thermal mutation analysis of the human-machine exoskeleton drive pre-stress is carried out, specifically including: directly extracting the temperature mutation compensation coefficient, long-term creep compensation coefficient, and attenuation term factor from the human-machine exoskeleton experimental calibration data in the human-machine exoskeleton collaborative multimodal signal fusion database; obtaining the average temperature change rate by averaging the array of temperature measurement sensors arranged at three axial points of the human-machine exoskeleton collaborative hydraulic joint shell; coupling analysis of the temperature mutation compensation coefficient and the average temperature change rate to obtain the human-machine exoskeleton collaborative environmental response component; comprehensively coupling analysis of the long-term creep compensation coefficient, the cumulative number of thermal cycles, the attenuation term factor, and the natural constant to obtain the human-machine exoskeleton collaborative long-term compensation component; obtaining the human-machine exoskeleton human-machine collaborative component through human-machine collaborative data coupling analysis; jointly coupling analysis of the human-machine exoskeleton collaborative environmental response component, the human-machine exoskeleton collaborative long-term compensation component, and the human-machine exoskeleton human-machine collaborative component to obtain the human-machine exoskeleton drive pre-stress current coefficient.

[0010] Furthermore, the human-machine collaborative component of the human-machine exoskeleton is obtained through the coupling analysis of human-machine collaborative data, specifically including: obtaining the human-machine electromyographic synergy coefficient through the joint coupling analysis of the coupling results of the human-machine electromyographic synergy coefficient and the corresponding weight factor and the coupling results of the human-machine action phase angle synergy coefficient and the corresponding weight factor; obtaining the human-machine collaborative component of the human-machine exoskeleton through the coupling analysis of the human-machine electromyographic synergy coefficient and the predefined human-machine impedance matching factor; the specific acquisition process of the human-machine electromyographic synergy coefficient is as follows: directly extracting the integral time window duration, muscle force safety lower limit threshold and muscle force attenuation coefficient from the human-machine exoskeleton experimental calibration data in the human-machine exoskeleton collaborative multimodal signal fusion database; obtaining the erector spinae electromyographic signal through real-time sampling of the human-machine exoskeleton textile electrode sports shirt; taking the envelope peak value in the actual standard box handling test of cold chain transportation The average value is used to obtain the initial maximum voluntary average contraction value; the muscle force safety component of the human-machine long-term muscle force dynamic correction component is obtained by coupling analysis of the initial maximum voluntary average contraction value and the muscle force safety lower limit threshold; the muscle force attenuation component of the human-machine long-term muscle force dynamic correction component is obtained by coupling analysis of the initial maximum voluntary average contraction value and the muscle force safety lower limit threshold; the muscle force safety component of the human-machine long-term muscle force dynamic correction component and the muscle force attenuation component of the human-machine long-term muscle force dynamic correction component are analyzed by the maximum value function to obtain the human-machine long-term muscle force dynamic correction component; the erector spinae muscle electromyographic signal is integrated according to the integration time window length and the collaborative detection time to obtain the human-machine electromyographic synergy integral basic value; the integration time window length, the human-machine electromyographic synergy integral basic value and the human-machine long-term muscle force dynamic correction component are coupled and analyzed to obtain the human-machine electromyographic synergy coefficient.

[0011] Furthermore, the specific process of obtaining the human-machine motion phase angle coordination coefficient is as follows: directly obtain the human-machine exoskeleton hip joint X-axis acceleration and human-machine exoskeleton hip joint Y-axis acceleration through the built-in control panel of the human-machine exoskeleton; analyze the human-machine exoskeleton hip joint X-axis acceleration and human-machine exoskeleton hip joint Y-axis acceleration through the inverse tangent function to obtain the human-machine motion phase angle coordination reference component; directly extract the temperature hysteresis coefficient, time adaptation constant, load coupling coefficient and load adaptation time constant from the human-machine exoskeleton collaborative multimodal signal fusion database; obtain the human-machine exoskeleton hip joint ambient temperature through sampling of the temperature measurement sensor on the outside of the human-machine exoskeleton hip joint; directly extract from the production log of the cold chain transportation refrigeration workshop The lowest extreme temperature of the cold chain transportation refrigeration environment and the human-machine exoskeleton load reference value are obtained; the action phase angle temperature compensation item is obtained through a comprehensive analysis of the temperature hysteresis coefficient, the human-machine exoskeleton hip joint environment temperature, the lowest extreme temperature of the cold chain transportation refrigeration environment, the collaborative detection time and the time adaptation constant; the real-time load is obtained according to the hydraulic pressure inversion through the built-in control panel of the human-machine exoskeleton; the action phase angle load compensation item is obtained through a comprehensive analysis of the load coupling coefficient, the real-time load, the human-machine exoskeleton load reference value and the load adaptation time constant; the human-machine action phase angle coordination coefficient is obtained through a comprehensive analysis of the human-machine action phase angle coordination reference component, the action phase angle temperature compensation item and the action phase angle load compensation item.

[0012] Furthermore, the human-machine collaborative component of the human-machine exoskeleton is obtained through the coupling analysis of human-machine collaborative data, which also includes: the weight factor corresponding to the human-machine electromyographic collaboration coefficient is the human-machine electromyographic collaboration coefficient weight factor; the weight factor of the human-machine motion phase angle collaboration coefficient is the human-machine motion phase angle collaboration weight factor; the specific acquisition process of the human-machine electromyographic collaboration coefficient weight factor is: directly extracting the human-machine electromyographic collaboration initial weight, electromyographic weight attenuation period and attenuation amplitude constant from the human-machine exoskeleton collaborative multimodal signal fusion database, and obtaining the human-machine electromyographic collaboration coefficient weight factor through the joint coupling analysis of the human-machine electromyographic collaboration initial weight, electromyographic weight attenuation period and attenuation amplitude constant; the specific acquisition process of the human-machine motion phase angle collaboration weight factor is: directly extracting the human-machine motion collaboration basic weight and angular velocity gain coefficient from the human-machine exoskeleton collaborative multimodal signal fusion database, directly collecting the human-machine exoskeleton hip joint angular velocity through the built-in control panel of the human-machine exoskeleton, and obtaining the human-machine motion phase angle collaboration weight factor through the joint coupling analysis of the human-machine motion collaboration basic weight, angular velocity gain coefficient and the human-machine exoskeleton hip joint angular velocity.

[0013] Furthermore, the pre-stress of the memory alloy spring is optimized and adjusted, specifically including: if the human-machine exoskeleton drive pre-stress current coefficient is less than the first human-machine exoskeleton drive pre-stress threshold, no additional adjustment is performed; if the human-machine exoskeleton drive pre-stress current coefficient is equal to or greater than the first human-machine exoskeleton drive pre-stress threshold, the human-machine exoskeleton drive pre-stress current is increased step by step until the human-machine exoskeleton drive pre-stress current coefficient is less than the first human-machine exoskeleton drive pre-stress threshold, and the step-by-step increase is stopped; if the human-machine exoskeleton drive pre-stress current is increased to the upper limit of the human-machine exoskeleton drive pre-stress current, the increase is stopped, and the time when the upper limit is reached is recorded as the limit time; if the limit time exceeds the warning time threshold, a warning is issued to notify relevant personnel.

[0014] Furthermore, myoelectric mutation analysis is performed to optimize and adjust the human-machine collaboration component of the human-machine exoskeleton or issue an early warning, specifically including: if the myoelectric synergy coefficient is greater than the myoelectric synergy coefficient threshold, no additional adjustment is performed; if the myoelectric synergy coefficient is less than or equal to the myoelectric synergy coefficient threshold, the real-time load fluctuation is judged. If the real-time load fluctuation value within the predefined mutation time is less than the real-time fluctuation load threshold, myoelectric gain compensation is performed. If the real-time load fluctuation value within the predefined mutation time is equal to or greater than the real-time fluctuation load threshold, a first-level early warning is initiated.

[0015] Furthermore, electromyographic gain compensation is performed, specifically including: directly extracting the low-temperature attenuation compensation coefficient and the historical average electromyographic value under the same working condition from the human-machine-exoskeleton collaborative multimodal signal fusion database; performing a proportion analysis of the historical average electromyographic value under the same working condition and the real-time erector spinae electromyographic signal, and then performing a comprehensive coupling analysis with the low-temperature attenuation compensation coefficient, the human-machine exoskeleton hip joint ambient temperature and the lowest value of the cold chain transportation refrigeration environment extreme temperature to obtain dynamic gain compensation; and updating the human-machine exoskeleton human-machine collaborative component according to the dynamic gain compensation.

[0016] Furthermore, a first-level warning is initiated, specifically including: gradually reducing the human-machine exoskeleton drive pre-pressure current until the human-machine exoskeleton drive pre-pressure current coefficient is less than or equal to the human-machine exoskeleton drive pre-pressure first threshold, and stopping the gradual reduction, and limiting the maximum driving force of the human-machine exoskeleton to within the normal working condition driving force threshold.

[0017] Furthermore, myoelectric gain compensation is performed, which also includes: if the human-machine exoskeleton human-machine collaboration component is less than the human-machine exoskeleton human-machine collaboration component threshold and the human-machine exoskeleton continuous operation time exceeds the human-machine exoskeleton continuous operation time threshold, an early warning is issued and relevant personnel are notified.

[0018] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0019] 1. Under actual cold chain temperature jump conditions, from -25°C to 5°C, a three-point axial temperature sensing array captures the sealing surface temperature gradient in real time, shortening the calculation delay of the environmental response component. Dynamically adjusting the preload pressure in conjunction with the temperature compensation coefficient addresses leakage issues caused by thermal expansion and contraction of the sealing material, reducing the leakage rate and thus improving the continuity of human-machine exoskeleton collaboration.

[0020] 2. To address the myoelectric attenuation caused by low-temperature muscle rigidity, textile electrode signals are fused with millimeter-wave radar motion capture, and a muscle force safety threshold mechanism is established to prevent misjudgment of fatigue operations. Dynamic gain is used to compensate for low-temperature signal loss, further improving the accuracy of action intention recognition, avoiding physical overload injuries to workers caused by power assistance lag, and improving the safety of human-machine exoskeleton collaboration.

[0021] 3. When myoelectric abnormalities are detected, the pre-pressure current is reduced to maintain basic sealing and limit the operating speed. This linkage control reduces the risk of system collapse under sudden load conditions and ensures safety in dangerous scenarios such as slippery ground. At the same time, through the creep compensation coefficient attenuation model, the aging correction of the compensation material is increased with each multiple hot and cold cycles, and the fatigue warning is triggered when the collaborative component exceeds the threshold, meeting the daily high-frequency in and out of the cold chain.

[0022] 4. By building a three-level protection system for current, temperature, and load, the system stability is improved under extreme working conditions, eliminating the risk of human-machine collaboration failure caused by hydraulic oil leakage. Through a four-order closed loop of environmental perception, signal fusion, dynamic decision-making, and safety protection, leakage control, action synchronization, and system reliability are achieved in cold chain scenarios, solving the problem of exoskeleton collaboration misalignment in extreme cold chain environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 Flowchart of the human-machine-exoskeleton collaboration method based on multimodal signal fusion provided in an embodiment of the present application. DETAILED DESCRIPTION

[0024] The embodiment of the present application provides a human-machine-exoskeleton collaboration method based on multimodal signal fusion.

[0025] In order to better understand the above technical solution, the above technical solution will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0026] like Figure 1As shown, it is a flow chart of the human-machine-exoskeleton collaboration method based on multimodal signal fusion provided in an embodiment of the present application, and the method includes the following steps: obtaining the human-machine-exoskeleton human-machine collaboration component through human-machine collaboration data coupling analysis; performing a cold and hot mutation analysis of the human-machine-exoskeleton drive pre-pressure in combination with the human-machine-exoskeleton human-machine collaboration component; performing electromyographic mutation analysis to optimize and adjust the human-machine-exoskeleton human-machine collaboration component or issue an early warning; and optimizing and adjusting the memory alloy spring pre-pressure in combination with the cold and hot mutation analysis results of the human-machine-exoskeleton drive pre-pressure.

[0027] Furthermore, a thermal mutation analysis of the human-machine exoskeleton drive pre-stress is carried out, specifically including: directly extracting the temperature mutation compensation coefficient, long-term creep compensation coefficient, and attenuation term factor from the human-machine exoskeleton experimental calibration data in the human-machine exoskeleton collaborative multimodal signal fusion database; obtaining the average temperature change rate by averaging the array of temperature measurement sensors arranged at three axial points of the human-machine exoskeleton collaborative hydraulic joint shell; coupling analysis of the temperature mutation compensation coefficient and the average temperature change rate to obtain the human-machine exoskeleton collaborative environmental response component; comprehensively coupling analysis of the long-term creep compensation coefficient, the cumulative number of thermal cycles, the attenuation term factor, and the natural constant to obtain the human-machine exoskeleton collaborative long-term compensation component; obtaining the human-machine exoskeleton human-machine collaborative component through human-machine collaborative data coupling analysis; jointly coupling analysis of the human-machine exoskeleton collaborative environmental response component, the human-machine exoskeleton collaborative long-term compensation component, and the human-machine exoskeleton human-machine collaborative component to obtain the human-machine exoskeleton drive pre-stress current coefficient.

[0028] In this embodiment, two-component gradient sealing rings are installed in the hydraulic cylinder piston rod sealing groove and the rotary joint end face sealing ring. For example, this sealing ring utilizes a layered composite structure: an inner layer of hydrogenated nitrile butadiene rubber (HNBR), which directly contacts the hydraulic fluid and provides a basic elastic seal; an outer layer of nickel-titanium shape memory alloy (NiTi-SMA) spring, pre-compressed at 20%, wrapped around the outer rubber layer. At -25°C, the HNBR maintains its flexibility, providing a basic seal. At this point, the SMA is in the martensite phase, providing rigid support and preventing low-temperature brittle cracking of the rubber. When the ambient temperature rises above 5°C, the SMA triggers an austenitic phase transformation. At this point, the spring contracts, generating a radial compensation pressure of 0.5 MPa, precisely offsetting the permanent deformation and relaxation of the rubber caused by repeated hot and cold cycling. This dual-state synergistic mechanism of rigid support at low temperatures and active compensation at elevated temperatures eliminates the alternating hardening and relaxation of the material caused by sudden temperature changes, thereby reducing the hydraulic oil leakage rate. This constitutes the hardware foundation for the software control described below.

[0029] The constraint of the pre-pressure current coefficient of the human-machine exoskeleton drive is as follows: ISMA=K1*|dT / dt|+K2*Nc ycle*e -R*Ncycle+λ*human(t), where ISMA represents the preload current coefficient for the human-machine exoskeleton drive; K1 represents the temperature jump compensation coefficient, extracted directly from human-machine exoskeleton experimental calibration data in the human-machine exoskeleton collaborative multimodal signal fusion database. For example, a linear temperature shock is applied to the sealing component in a temperature-controlled chamber between -25°C and 25°C, with slopes set to 1°C / s, 2°C / s, and 5°C / s. The relationship between the SMA current and hydraulic leakage at different slopes is measured. The data is fitted using the least squares method to the average slope, resulting in K1 = 0.25.

[0030] dT / dt represents the average temperature change rate, calculated by averaging the PT1000 array arranged at three axial points on the human-machine exoskeleton's hydraulic joint housing. N_cycle represents the cumulative number of hot and cold cycles, collected by the exoskeleton's built-in control board using a temperature jump counter. e represents a natural constant, and R represents the attenuation factor, extracted directly from experimental calibration data in the human-machine exoskeleton collaborative multimodal signal fusion database. For example, when fitting the HNBR aging curve, if the compensation decreases after the cumulative number of hot and cold cycles exceeds 500, then R = 1 / 500 = 0.002. K2 represents the long-term creep compensation coefficient, extracted directly from experimental calibration data in the human-machine exoskeleton collaborative multimodal signal fusion database. For example, through accelerated aging experiments, HNBR test pieces are placed in a 120°C oven, and the test pieces are accelerated 50 times compared to normal temperature cycles. The test pieces are taken out every 24 hours to measure the compression set rate. The long-term creep compensation coefficient is obtained by fitting the relationship curve between the cumulative number of hot and cold cycles and the current compensation amount.

[0031] ΔP_hyd represents the pressure differential across the seal, which is acquired by the exoskeleton's built-in control board using piezoresistive differential pressure sensors at the hydraulic cylinder inlet and outlet. λ represents the predefined human-machine impedance matching factor, extracted directly from calibration data from human-machine exoskeleton experiments within the human-machine exoskeleton collaborative multimodal signal fusion database. This is determined through human-machine impedance matching experiments. For example, a subject wearing the exoskeleton repeatedly carries objects while simultaneously collecting the lumbar spine load Fspine and the machine's hydraulic output Fhyd. The human-machine force error constraint is formulated as min∑[(Fspine - Fhyd) / Fspine]. λ is essentially the trust weight of the human state in the machine's regulation, ranging from 0 (completely ignoring human signals) to 1 (completely relying on human signals). The minimum human-machine force error is obtained by fitting the human-machine force error under different human-machine impedance matching factors. The corresponding human-machine impedance matching factor is the predefined human-machine impedance matching factor. human(t) represents the human-machine synergy coefficient, obtained through coupled analysis of human-machine collaborative data.

[0032] It should be noted that t represents the collaborative detection moment, which is used to quantify the time interval from the start of timing to the collaborative detection moment. The above human-machine exoskeleton drive pre-pressure current coefficient constraint publicizes other parameters besides the human-machine collaborative coefficient, which also represent the parameters at the collaborative detection moment.

[0033] Furthermore, the human-machine collaborative component of the human-machine exoskeleton is obtained through the coupling analysis of human-machine collaborative data, specifically including: obtaining the human-machine electromyographic synergy coefficient through the joint coupling analysis of the coupling results of the human-machine electromyographic synergy coefficient and the corresponding weight factor and the coupling results of the human-machine action phase angle synergy coefficient and the corresponding weight factor; obtaining the human-machine collaborative component of the human-machine exoskeleton through the coupling analysis of the human-machine electromyographic synergy coefficient and the predefined human-machine impedance matching factor; the specific acquisition process of the human-machine electromyographic synergy coefficient is as follows: directly extracting the integral time window duration, muscle force safety lower limit threshold and muscle force attenuation coefficient from the human-machine exoskeleton experimental calibration data in the human-machine exoskeleton collaborative multimodal signal fusion database; obtaining the erector spinae electromyographic signal through real-time sampling of the human-machine exoskeleton textile electrode sports shirt; taking the envelope peak value in the actual standard box handling test of cold chain transportation The average value is used to obtain the initial maximum voluntary average contraction value; the muscle force safety component of the human-machine long-term muscle force dynamic correction component is obtained by coupling analysis of the initial maximum voluntary average contraction value and the muscle force safety lower limit threshold; the muscle force attenuation component of the human-machine long-term muscle force dynamic correction component is obtained by coupling analysis of the initial maximum voluntary average contraction value and the muscle force safety lower limit threshold; the muscle force safety component of the human-machine long-term muscle force dynamic correction component and the muscle force attenuation component of the human-machine long-term muscle force dynamic correction component are analyzed by the maximum value function to obtain the human-machine long-term muscle force dynamic correction component; the erector spinae muscle electromyographic signal is integrated according to the integration time window length and the collaborative detection time to obtain the human-machine electromyographic synergy integral basic value; the integration time window length, the human-machine electromyographic synergy integral basic value and the human-machine long-term muscle force dynamic correction component are coupled and analyzed to obtain the human-machine electromyographic synergy coefficient.

[0034] In this embodiment, human(t) = α(t)*Kmuscle(t) + β(t)*sin(φmotion(t)), where α(t) represents the human-machine myoelectric synergy coefficient weighting factor, and β(t) represents the human-machine motion phase angle synergy weighting factor. sin() converts physical angles into biomechanical characteristic quantities. Human joint motion is essentially a periodic function, which is used to match the sinusoidal characteristics of joint torque and avoid angle discontinuities.

[0035] The constraint formula of human-machine myoelectric synergy coefficient is as follows: Among them, Kmuscle(t) represents the myoelectric synergy coefficient, and TW represents the integration time window length, which is directly extracted from the human-machine-exoskeleton collaborative multimodal signal fusion database. For example, it is set to 0.5 based on the motor nerve conduction delay of the human body.

[0036] SEMG(t) represents the real-time erector spinae muscle electromyographic signal, which is obtained by real-time sampling of the human-machine exoskeleton textile electrode sweatshirt with a sampling rate of 1 kHz.

[0037] MVCadj(t) represents the dynamic correction component of the long-term muscle force of the human machine, MVCadj(t)=max(MVC0*e- λm(t) ,XS1*MVC0), where XS1 represents the lower safety threshold of muscle force, which is directly extracted from the human-machine-exoskeleton collaborative multimodal signal fusion database. According to the expert prior knowledge, the risk of loss of control of the movement increases sharply when the muscle force is less than 40% MVC, so XS1 is set to 0.4.

[0038] MVC0 represents the initial maximum autonomous mean shrinkage value, which is obtained by taking the average value of the envelope peak value in the actual standard box handling test of cold chain transportation.

[0039] λm(t) represents the muscle force attenuation coefficient, which is directly extracted from the human-machine-exoskeleton experimental calibration data in the human-machine-exoskeleton collaborative multimodal signal fusion database. For example, the muscle force attenuation fitting coefficient is obtained by fitting the fatigue experiment to obtain the muscle force attenuation coefficient.

[0040] Through dynamic correction, muscle fatigue caused by working time is taken into account, making the normalization of electromyographic signals more reasonable.

[0041] Furthermore, the specific process of obtaining the human-machine motion phase angle coordination coefficient is as follows: directly obtain the human-machine exoskeleton hip joint X-axis acceleration and human-machine exoskeleton hip joint Y-axis acceleration through the built-in control panel of the human-machine exoskeleton; analyze the human-machine exoskeleton hip joint X-axis acceleration and human-machine exoskeleton hip joint Y-axis acceleration through the inverse tangent function to obtain the human-machine motion phase angle coordination reference component; directly extract the temperature hysteresis coefficient, time adaptation constant, load coupling coefficient and load adaptation time constant from the human-machine exoskeleton collaborative multimodal signal fusion database; obtain the human-machine exoskeleton hip joint ambient temperature through sampling of the temperature measurement sensor on the outside of the human-machine exoskeleton hip joint; directly extract from the production log of the cold chain transportation refrigeration workshop The lowest extreme temperature of the cold chain transportation refrigeration environment and the human-machine exoskeleton load reference value are obtained; the action phase angle temperature compensation item is obtained through a comprehensive analysis of the temperature hysteresis coefficient, the human-machine exoskeleton hip joint environment temperature, the lowest extreme temperature of the cold chain transportation refrigeration environment, the collaborative detection time and the time adaptation constant; the real-time load is obtained according to the hydraulic pressure inversion through the built-in control panel of the human-machine exoskeleton; the action phase angle load compensation item is obtained through a comprehensive analysis of the load coupling coefficient, the real-time load, the human-machine exoskeleton load reference value and the load adaptation time constant; the human-machine action phase angle coordination coefficient is obtained through a comprehensive analysis of the human-machine action phase angle coordination reference component, the action phase angle temperature compensation item and the action phase angle load compensation item.

[0042] In this embodiment, the human-machine motion phase angle coordination coefficient constraint formula is as follows: in, Indicates the X-axis acceleration of the human-machine exoskeleton hip joint, which is directly obtained through the built-in control panel of the human-machine exoskeleton. The Y-axis acceleration of the human-machine exoskeleton hip joint is obtained directly from the built-in control panel of the human-machine exoskeleton. Δφtemp(t) represents the temperature compensation term for the motion phase angle, and Δφload(t) represents the load compensation term for the motion phase angle. The two-dimensional acceleration of the hip joint is converted to spatial angles using arctan²(,) for basic kinematic transformation.

[0043] Δφtemp(t) = kT*(Tamb+EDLL)*sigmoid(t / CST), where kT represents the temperature hysteresis coefficient, which is directly extracted from human-machine-exoskeleton experimental calibration data in the human-machine-exoskeleton collaborative multimodal signal fusion database. For example, a low-temperature gait experiment was used for calibration, with a ramp temperature increase from -25°C to 5°C. The changes in the human-machine-exoskeleton driving force were measured at different temperatures. The temperature hysteresis coefficients at different temperatures were obtained, with the human-machine-exoskeleton driving force at 5°C as a unit of 1. Tamb represents the ambient temperature of the human-machine-exoskeleton hip joint, which is sampled using a PT1000 platinum resistor installed on the outside of the hip joint. EDLL represents the lowest extreme temperature of the cold chain transport refrigeration environment, which is directly extracted from the production log of the cold chain transport refrigeration workshop. For example, the glass transition temperature of hydrogenated nitrile butadiene rubber (HNBR) is approximately -25°C, below which the material's elasticity drops sharply. Using -25°C as the reference point to match the material's critical phase transition temperature, the EDLL is -25. CST represents the time adaptation constant, which is directly extracted from the historical data of the human-machine-exoskeleton in the human-machine-exoskeleton collaborative multimodal signal fusion database. It is used to represent the non-carrying time of human-machine collaboration when entering and leaving the cold chain transport refrigeration workshop, in addition to the formal handling of goods. It is obtained by the average value in the historical data.

[0044] Δφload(t)=cL*(Fload(t) / FZJZ)*(1-e -t / WDCS ), where cL represents the load coupling coefficient, which is extracted from the experimental calibration data in the human-machine-exoskeleton collaborative multimodal signal fusion database. For example, the hip angle offset is measured by optical motion capture under different loads through a step load experiment to obtain the fitting coefficient, which is the load coupling coefficient; Fload(t) represents the real-time load, which is obtained by inverting the hydraulic pressure through the built-in control panel of the human-machine exoskeleton, for example: Fload = piston area ratio × hydraulic cylinder pressure; FZJZ represents the human-machine exoskeleton load baseline value, which is directly extracted from the human-machine exoskeleton production log and is used to represent the load that the human-machine exoskeleton can withstand during normal operation.

[0045] WDCS stands for load adaptation time constant, which is directly extracted from the human-machine-exoskeleton collaborative multimodal signal fusion database and obtained based on expert prior knowledge. It is used to describe the time required from the start of load handling to the dynamic equilibrium of the muscle discharge pattern. Specifically, it can be set to when the fluctuation rate of the electromyographic amplitude envelope is <5%, it is judged as stable.

[0046] Furthermore, the human-machine collaborative component of the human-machine exoskeleton is obtained through the coupling analysis of human-machine collaborative data, which also includes: the weight factor corresponding to the human-machine electromyographic collaboration coefficient is the human-machine electromyographic collaboration coefficient weight factor; the weight factor of the human-machine motion phase angle collaboration coefficient is the human-machine motion phase angle collaboration weight factor; the specific acquisition process of the human-machine electromyographic collaboration coefficient weight factor is: directly extracting the human-machine electromyographic collaboration initial weight, electromyographic weight attenuation period and attenuation amplitude constant from the human-machine exoskeleton collaborative multimodal signal fusion database, and obtaining the human-machine electromyographic collaboration coefficient weight factor through the joint coupling analysis of the human-machine electromyographic collaboration initial weight, electromyographic weight attenuation period and attenuation amplitude constant; the specific acquisition process of the human-machine motion phase angle collaboration weight factor is: directly extracting the human-machine motion collaboration basic weight and angular velocity gain coefficient from the human-machine exoskeleton collaborative multimodal signal fusion database, directly collecting the human-machine exoskeleton hip joint angular velocity through the built-in control panel of the human-machine exoskeleton, and obtaining the human-machine motion phase angle collaboration weight factor through the joint coupling analysis of the human-machine motion collaboration basic weight, angular velocity gain coefficient and the human-machine exoskeleton hip joint angular velocity.

[0047] In this embodiment, α(t) represents the weight factor of the human-machine myoelectric synergy coefficient, α(t) = α0*[1-SJFD*(t / THDJ)], where α0 represents the initial weight of human-machine myoelectric synergy, which is directly extracted from the human-machine-exoskeleton collaborative multimodal signal fusion database and is generally set to 0.5. THDJ represents the myoelectric weight attenuation period, which is directly extracted from the human-machine-exoskeleton collaborative multimodal signal fusion database and is specifically obtained based on expert prior knowledge and is generally set to one hour. SJFD represents the attenuation amplitude constant, which is directly extracted from the experimental calibration in the human-machine-exoskeleton collaborative multimodal signal fusion database. For example, in the later stage of the operation, the myoelectric noise gradually increases, and the weight needs to be reduced accordingly. The working time and myoelectric noise are fitted according to the experimental data to obtain the corresponding attenuation amplitude constant.

[0048] β(t) represents the human-machine motion phase angle coordination weight factor, β(t) = β0 + Δβ*|dφ / dt|, β0 represents the basic weight of human-machine motion coordination, which is directly extracted from the human-machine-exoskeleton collaborative multimodal signal fusion database and is generally taken as 0.5. Δβ represents the angular velocity gain coefficient, which is directly extracted from the calibration data in the human-machine-exoskeleton collaborative multimodal signal fusion database. For example, in the calibration experiment, the weight increases by 0.008 for every 1° / s increase in angular velocity, which is used to match the needs of rapid obstacle avoidance. dφ / dt represents the angular velocity of the human-machine exoskeleton hip joint, which is directly collected through the built-in control panel of the human-machine exoskeleton.

[0049] Furthermore, the pre-stress of the memory alloy spring is optimized and adjusted, specifically including: if the human-machine exoskeleton drive pre-stress current coefficient is less than the first human-machine exoskeleton drive pre-stress threshold, no additional adjustment is performed; if the human-machine exoskeleton drive pre-stress current coefficient is equal to or greater than the first human-machine exoskeleton drive pre-stress threshold, the human-machine exoskeleton drive pre-stress current is increased step by step until the human-machine exoskeleton drive pre-stress current coefficient is less than the first human-machine exoskeleton drive pre-stress threshold, and the step-by-step increase is stopped; if the human-machine exoskeleton drive pre-stress current is increased to the upper limit of the human-machine exoskeleton drive pre-stress current, the increase is stopped, and the time when the upper limit is reached is recorded as the limit time; if the limit time exceeds the warning time threshold, a warning is issued to notify relevant personnel.

[0050] Furthermore, myoelectric mutation analysis is performed to optimize and adjust the human-machine collaboration component of the human-machine exoskeleton or issue an early warning, specifically including: if the myoelectric synergy coefficient is greater than the myoelectric synergy coefficient threshold, no additional adjustment is performed; if the myoelectric synergy coefficient is less than or equal to the myoelectric synergy coefficient threshold, the real-time load fluctuation is judged. If the real-time load fluctuation value within the predefined mutation time is less than the real-time fluctuation load threshold, myoelectric gain compensation is performed. If the real-time load fluctuation value within the predefined mutation time is equal to or greater than the real-time fluctuation load threshold, a first-level early warning is initiated.

[0051] In this embodiment, when the electromyographic signal is abnormal but the joint movement is normal, it may indicate signal acquisition interference rather than actual movement disorder. In this case, signal gain compensation is used to restore control accuracy.

[0052] Furthermore, electromyographic gain compensation is performed, specifically including: directly extracting the low-temperature attenuation compensation coefficient and the historical average electromyographic value under the same working condition from the human-machine-exoskeleton collaborative multimodal signal fusion database; performing a proportion analysis of the historical average electromyographic value under the same working condition and the real-time erector spinae electromyographic signal, and then performing a comprehensive coupling analysis with the low-temperature attenuation compensation coefficient, the human-machine exoskeleton hip joint ambient temperature and the lowest value of the cold chain transportation refrigeration environment extreme temperature to obtain dynamic gain compensation; and updating the human-machine exoskeleton human-machine collaborative component according to the dynamic gain compensation.

[0053] In this embodiment, the dynamic gain compensation constraint is expressed as follows: Kcomp = (SEMGref / SEMG(t)) × [1 + DWSB × (Tamb + EDLL)], where Tamb represents the ambient temperature of the human-machine exoskeleton hip joint, EDLL represents the lowest extreme temperature of the cold chain transport refrigeration environment, SEMGref represents the historical average electromyographic value under the same working conditions, directly extracted from historical data in the human-machine exoskeleton collaborative multimodal signal fusion database, and DWSB represents the low-temperature attenuation compensation coefficient, directly extracted from the human-machine exoskeleton collaborative multimodal signal fusion database. For example, the erector spinae electromyographic value and ambient temperature are obtained in real time. After the ambient temperature drops, the maximum deviation between the measured and actual values ​​of the erector spinae electromyographic value and the ambient temperature are obtained. The low-temperature attenuation compensation coefficient is obtained by fitting the ratio of the two. SEMG(t) represents the real-time erector spinae electromyographic signal.

[0054] human(t)new = human(t) × Kcomp; human(t)new represents the updated human-machine cooperation coefficient, which in turn updates the human-machine cooperation component of the human-machine exoskeleton. Safety boundary control: the compensation upper limit Kcomp ≤ KcompYZ1, KcompYZ1 represents the dynamic gain compensation threshold to prevent excessive amplification of noise. When the ambient humidity is > 85%, Kcomp is reduced by an additional 20% to suppress sweat conductive interference.

[0055] It should be noted that the value in the human-machine exoskeleton drive pre-pressure current coefficient is human(t)ne w, which is the value obtained after adjustment.

[0056] Furthermore, a first-level warning is initiated, specifically including: gradually reducing the human-machine exoskeleton drive pre-pressure current until the human-machine exoskeleton drive pre-pressure current coefficient is less than or equal to the human-machine exoskeleton drive pre-pressure first threshold, and stopping the gradual reduction, and limiting the maximum driving force of the human-machine exoskeleton to within the normal working condition driving force threshold.

[0057] Furthermore, myoelectric gain compensation is performed, which also includes: if the human-machine exoskeleton human-machine collaboration component is less than the human-machine exoskeleton human-machine collaboration component threshold and the human-machine exoskeleton continuous operation time exceeds the human-machine exoskeleton continuous operation time threshold, an early warning is issued and relevant personnel are notified.

[0058] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0059] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0060] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0061] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0062] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0063] Obviously, those skilled in the art may make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if such modifications and variations fall within the scope of the claims and their equivalents, the present invention is intended to include such modifications and variations.

Claims

1. A human-machine exoskeleton collaboration method based on multimodal signal fusion, characterized in that: The following steps are involved: The human-machine collaboration component of the human-machine exoskeleton is obtained through the coupling analysis of human-machine collaboration data; Combined with the human-machine collaboration component of the human-machine exoskeleton, the cold and hot mutation analysis of the human-machine exoskeleton driving pre-stress was carried out; Conduct electromyographic mutation analysis to optimize and adjust the human-machine collaboration component of the human-machine exoskeleton or issue an early warning; The preload of the memory alloy spring is optimized and adjusted based on the analysis results of hot and cold mutation of the preload of the human-machine exoskeleton drive.

2. The human-machine exoskeleton collaboration method based on multimodal signal fusion according to claim 1, characterized in that: The analysis of thermal shock of pre-stress of the human-machine exoskeleton driving specifically includes: The temperature mutation compensation coefficient, long-term creep compensation coefficient, and attenuation factor are directly extracted from the human-machine exoskeleton experimental calibration data in the human-machine exoskeleton collaborative multimodal signal fusion database; The average temperature change rate is obtained by averaging the temperature sensor array arranged at three axial points of the human-machine exoskeleton collaborative hydraulic joint shell. The temperature mutation compensation coefficient and the average rate of temperature change are coupled and analyzed to obtain the environmental response component of the human-machine-exoskeleton collaboration. The long-term creep compensation coefficient, the cumulative number of hot and cold cycles, the attenuation factor and the natural constant are comprehensively coupled and analyzed to obtain the long-term compensation component of the human-machine exoskeleton collaboration. The human-machine collaboration component of the human-machine exoskeleton is obtained through the coupling analysis of human-machine collaboration data; The human-machine-exoskeleton collaborative environmental response component, the human-machine-exoskeleton collaborative long-term compensation component and the human-machine-exoskeleton human-machine collaborative component are coupled and analyzed together to obtain the human-machine-exoskeleton driving pre-pressure current coefficient.

3. The human-machine exoskeleton collaboration method based on multimodal signal fusion as claimed in claim 2, characterized in that: The human-machine collaboration component of the human-machine exoskeleton obtained through human-machine collaboration data coupling analysis specifically includes: The human-machine myoelectric synergy coefficient is obtained by jointly coupling the coupling results of the human-machine electromyographic synergy coefficient and the corresponding weight factor, and the coupling results of the human-machine action phase angle synergy coefficient and the corresponding weight factor. The human-machine synergy component of the human-machine exoskeleton is obtained by coupling the human-machine electromyographic synergy coefficient with the predefined human-machine impedance matching factor; The specific process of obtaining the human-machine electromyographic synergy coefficient is as follows: The integration time window duration, muscle force safety lower limit threshold and muscle force attenuation coefficient are directly extracted from the human-machine-exoskeleton experimental calibration data in the human-machine-exoskeleton collaborative multimodal signal fusion database; The erector spinae muscle electromyographic signals were collected in real time through the human-machine exoskeleton textile electrode sweatshirt; The initial maximum autonomous average shrinkage value is obtained by taking the average value of the envelope peak value in the actual standard box handling test of cold chain transportation; Through the coupling analysis of the initial maximum voluntary average contraction value and the lower limit threshold of muscle force safety, the dynamic correction component of the human-machine long-term muscle force and the muscle force safety component are obtained; Through the coupling analysis of the initial maximum voluntary average contraction value and the lower limit threshold of muscle force safety, the dynamic correction component of the human-machine long-term muscle force and the muscle force attenuation component are obtained; The muscle safety component and the muscle attenuation component of the long-term dynamic correction component of the human-machine muscle force are analyzed by the maximum value function to obtain the long-term dynamic correction component of the human-machine muscle force; The erector spinae muscle electromyographic signal is integrated according to the integration time window length and the collaborative detection time to obtain the human-machine electromyographic collaborative integral basic value; The integration time window duration, the basic value of human-machine myoelectric synergy integral and the dynamic correction component of human-machine long-term muscle force were coupled and analyzed to obtain the human-machine myoelectric synergy coefficient.

4. The human-machine exoskeleton collaboration method based on multimodal signal fusion according to claim 3, characterized in that: The specific process of obtaining the human-machine action phase angle coordination coefficient is as follows: The X-axis acceleration and Y-axis acceleration of the hip joint of the human-machine exoskeleton are directly obtained through the built-in control panel of the human-machine exoskeleton; The X-axis acceleration and Y-axis acceleration of the hip joint of the human-machine exoskeleton are analyzed by the inverse tangent function to obtain the human-machine motion phase angle coordination reference component. The temperature hysteresis coefficient, time adaptation constant, load coupling coefficient and load adaptation time constant are directly extracted from the human-machine-exoskeleton collaborative multimodal signal fusion database; The ambient temperature of the exoskeleton hip joint is obtained through sampling by the temperature sensor on the outside of the exoskeleton hip joint; Directly extracted from the production log of the cold chain transport refrigeration workshop, the lowest extreme temperature value of the cold chain transport refrigeration environment and the human-machine exoskeleton load benchmark value; The temperature compensation term of the motion phase angle is obtained through comprehensive analysis of the temperature hysteresis coefficient, the ambient temperature of the human-machine exoskeleton hip joint, the lowest extreme temperature of the cold chain transportation refrigeration environment, the collaborative detection time and the time adaptation constant. The real-time load is obtained by inverting the hydraulic pressure through the built-in control panel of the human-machine exoskeleton; The motion phase angle load compensation term is obtained through comprehensive analysis of the load coupling coefficient, real-time load, human-machine exoskeleton load baseline value, and load adaptation time constant. The human-machine motion phase angle synergy coefficient is obtained by comprehensively analyzing the human-machine motion phase angle synergy benchmark component, motion phase angle temperature compensation term and motion phase angle load compensation term.

5. The human-machine exoskeleton collaboration method based on multimodal signal fusion as claimed in claim 3, characterized in that: The method of obtaining the human-machine-exoskeleton human-machine collaboration component through human-machine collaboration data coupling analysis also includes: The weight factor corresponding to the human-machine myoelectric synergy coefficient is the human-machine myoelectric synergy coefficient weight factor; The weight factor of the human-machine action phase angle coordination coefficient is the human-machine action phase angle coordination weight factor; The specific process of obtaining the human-machine myoelectric synergy coefficient weight factor is as follows: The initial weight, attenuation period, and attenuation amplitude constant of human-machine electromyography are directly extracted from the human-machine exoskeleton collaborative multimodal signal fusion database. The weight factor of the human-machine electromyography synergy coefficient is obtained through the joint coupling analysis of the initial weight, attenuation period, and attenuation amplitude constant of human-machine electromyography synergy. The specific process of obtaining the human-machine action phase angle coordination weight factor is as follows: The basic weights and angular velocity gain coefficients of human-machine motion collaboration are directly extracted from the human-machine-exoskeleton collaborative multimodal signal fusion database. The angular velocity of the human-machine-exoskeleton hip joint is directly collected through the built-in control panel of the human-machine-exoskeleton. The human-machine motion phase angle collaboration weight factor is obtained through the joint coupling analysis of the basic weights, angular velocity gain coefficients and the angular velocity of the human-machine-exoskeleton hip joint.

6. The human-machine exoskeleton collaboration method based on multimodal signal fusion as claimed in claim 2, characterized in that: The optimizing and adjusting the preload of the memory alloy spring specifically includes: If the exoskeleton driving pre-pressure current coefficient is less than the first exoskeleton driving pre-pressure threshold, no additional adjustment is performed; If the human-machine exoskeleton drive pre-pressure current coefficient is equal to or greater than the human-machine exoskeleton drive pre-pressure first threshold, the human-machine exoskeleton drive pre-pressure current is increased step by step until the human-machine exoskeleton drive pre-pressure current coefficient is less than the human-machine exoskeleton drive pre-pressure first threshold, and then the step-by-step increase is stopped. If the human-machine exoskeleton drive pre-pressure current is increased to the human-machine exoskeleton drive pre-pressure current upper limit, the increase is stopped, and the time when the upper limit is reached is recorded as the limit time. If the limit time exceeds the warning time threshold, a warning is issued to notify relevant personnel.

7. The human-machine exoskeleton collaboration method based on multimodal signal fusion according to claim 1, characterized in that: The myoelectric mutation analysis is performed to optimize and adjust the human-machine collaboration component of the human-machine exoskeleton or issue an early warning, specifically including: If the EMG synergy coefficient is greater than the EMG synergy coefficient threshold, no additional adjustment is performed; If the myoelectric synergy coefficient is less than or equal to the myoelectric synergy coefficient threshold, the real-time load fluctuation is judged. If the real-time load fluctuation value within the predefined mutation time is less than the real-time fluctuation load threshold, myoelectric gain compensation is performed. If the real-time load fluctuation value within the predefined mutation time is equal to or greater than the real-time fluctuation load threshold, a first-level warning is initiated.

8. The human-machine exoskeleton collaboration method based on multimodal signal fusion according to claim 7, characterized in that: The electromyography gain compensation specifically includes: The low-temperature attenuation compensation coefficient and the historical average electromyographic value under the same working conditions are directly extracted from the human-machine-exoskeleton collaborative multimodal signal fusion database; The historical average EMG value and real-time EMG signal of the erector spinae muscles under the same working conditions are analyzed for their proportions. This is then coupled with the low-temperature attenuation compensation coefficient, the ambient temperature of the human-machine exoskeleton hip joint, and the lowest extreme temperature of the cold chain transportation refrigeration environment to obtain dynamic gain compensation. Update the human-machine collaboration component of the human-machine exoskeleton based on dynamic gain compensation.

9. The human-machine exoskeleton collaboration method based on multimodal signal fusion according to claim 7, characterized in that: The initiation of the first-level warning specifically includes: The exoskeleton driving pre-pressure current is gradually reduced until the exoskeleton driving pre-pressure current coefficient is less than or equal to the exoskeleton driving pre-pressure first threshold, and the reduction is stopped step by step, and the maximum driving force of the exoskeleton is limited to within the normal working condition driving force threshold.

10. The human-machine exoskeleton collaboration method based on multimodal signal fusion according to claim 8, characterized in that: The electromyography gain compensation further includes: If the human-machine exoskeleton human-machine collaboration component is less than the human-machine exoskeleton human-machine collaboration component threshold and the human-machine exoskeleton continuous operation time exceeds the human-machine exoskeleton continuous operation time threshold, an early warning will be issued and relevant personnel will be notified.

Images