A control method and system for an exoskeleton robot

Abstract

The application relates to a control method and system of an exoskeleton robot. The method comprises the following steps: calculating joint interaction power according to the correlation between joint angular velocity and human-machine interaction torque of the exoskeleton robot in a target period; analyzing the distribution of the amplitudes of a plurality of historical joint interaction powers to obtain fluctuation characteristic values; performing gait analysis on a user in the target period to obtain a target gait stage in which the target period is located, and determining a force upper limit corresponding to the target period; when the fluctuation characteristic values of the exoskeleton robot in the target period exceed a preset fluctuation threshold, the human-machine interaction state of the target period is a reverse force state, and the amplitude of the joint interaction power of the target period exceeds the corresponding force upper limit, it is determined that the target period is in an emergency braking state, and a stiffness adjustment amount of the target period is calculated based on the fluctuation characteristic values of the target period. The application can improve the recognition accuracy of the emergency braking action of the exoskeleton robot.

Description

Technical Field

[0001] This application relates to the field of robot control technology, specifically to a control method and system for an exoskeleton robot. Background Technology

[0002] Exoskeleton robots are widely used as auxiliary load-bearing equipment in industrial logistics and outdoor operations. In actual operations, workers often face unstructured environments such as gravel roads or irregular steps. In such environments, wearers frequently need to make sudden stops or take small steps to avoid obstacles or missteps, causing the muscles of the lower limbs to contract instantaneously and generate a backward braking torque. To ensure safety, it is necessary to directly identify sudden force changes at the underlying microcontroller. Existing underlying data processing methods mainly face the following technical contradictions: if a long-window low-pass filter is added to the microcontroller to smooth the force data, it will cause a decision delay of tens to hundreds of milliseconds, missing the best safe time to release the torque; if the first-order difference or derivative is used to calculate the rate of change of force, it will proportionally amplify the high-frequency mechanical vibration and sensor electrical noise caused by motor operation, causing the system to frequently stop erroneously during normal walking. Summary of the Invention

[0003] To address the aforementioned technical problems, the purpose of this application is to provide a control method and system for an exoskeleton robot, the specific technical solution of which is as follows: Firstly, a control method for an exoskeleton robot is provided, the method comprising: Based on the correlation between the joint angular velocity and human-machine interaction torque of the exoskeleton robot in the target cycle, the joint interaction power of the target cycle is calculated; whereby the human-machine interaction torque characterizes the interaction torque between the exoskeleton robot and the human body, and the joint interaction power is used to characterize the intensity of the human-machine interaction. The distribution of the amplitude of the interaction power of multiple historical joints before the target cycle is analyzed to obtain the fluctuation characteristic value of the exoskeleton robot in the target cycle; the fluctuation characteristic value characterizes the degree of force waveform distortion of the exoskeleton robot in the target cycle. Gait analysis is performed on the user during the target cycle to obtain the target gait stage of the target cycle, and the upper limit of force corresponding to the target cycle is determined based on the target gait stage. When the fluctuation characteristic value of the exoskeleton robot in the target cycle exceeds the preset fluctuation threshold, the human-machine interaction state in the target cycle is in the reverse force state, and the amplitude of the joint interaction power in the target cycle exceeds the corresponding force limit, it is determined that the target cycle is in an emergency braking state, and the stiffness adjustment amount of the target cycle is calculated based on the fluctuation characteristic value of the target cycle; wherein, the human-machine interaction state is determined according to the sign of the joint interaction power, and the stiffness adjustment amount is used to reduce the output stiffness of the exoskeleton robot joint.

[0004] Optionally, based on the correlation between the joint angular velocity and human-robot interaction torque of the exoskeleton robot in the target cycle, the joint interaction power of the target cycle is calculated, including: During the target cycle, the angular positions of the joints of the exoskeleton robot are collected, and the angular positions of adjacent cycles are differentially calculated to obtain the joint angular velocity of the target cycle. During the target cycle, the torque of human-machine interaction on the exoskeleton robot is collected by a torque sensor; The joint interaction power of the exoskeleton robot during the target cycle is obtained by multiplying the joint angular velocity and the human-machine interaction torque.

[0005] Optionally, the distribution of amplitudes of the interaction power of multiple historical joints prior to the target cycle is analyzed to obtain the fluctuation characteristic values ​​of the exoskeleton robot during the target cycle, including: Based on the degree of deviation between the amplitudes of the interaction power of multiple historical joints before the target cycle, the fluctuation characteristic value of the exoskeleton robot in the target cycle is calculated.

[0006] Optionally, based on the degree of deviation between the amplitudes of the interaction power of multiple historical joints prior to the target cycle, the fluctuation characteristic values ​​of the exoskeleton robot in the target cycle are calculated, including: Calculate the square of the amplitude of the historical joint interaction power of the previous cycle of the target cycle to obtain the square value of the target cycle; The product of the amplitude of the joint interaction power in the target period and the amplitude of the historical joint interaction power in the period two years prior to the target period is calculated to obtain the product value of the target period. The difference between the square of the target period and its product is used to obtain the fluctuation characteristic value of the target period.

[0007] Optionally, gait analysis is performed on the user during the target period to obtain the target gait stage of the target period, including: During the target cycle, gait phase values ​​are calculated based on user motion data; wherein, the gait phase values ​​are percentage values ​​used to characterize the phase of the gait cycle; The gait phase value is determined as the target gait phase in which the target cycle is located; wherein, the target gait phase is the phase percentage obtained by performing phase localization of human gait.

[0008] Optionally, the upper limit of force corresponding to the target cycle is determined based on the target gait stage, including: Using the target gait stage of the target cycle as an index, a query is performed in a preset reference mapping table to extract the upper limit of force corresponding to the target gait stage, thus obtaining the upper limit of force corresponding to the target cycle; the upper limit of force corresponding to each stage in the human gait cycle is stored in the reference mapping table in advance.

[0009] Optionally, when the fluctuation characteristic value of the exoskeleton robot in the target cycle exceeds a preset fluctuation threshold, the human-machine interaction state in the target cycle is in a reverse force state, and the amplitude of the joint interaction power in the target cycle exceeds the corresponding force limit, it is determined that the target cycle is in an emergency braking state, and the stiffness adjustment amount of the target cycle is calculated based on the fluctuation characteristic value of the target cycle, including: When the joint interaction power of the target cycle is negative, the human-machine interaction state of the target cycle is determined to be the reverse force state. When the fluctuation characteristic value of the exoskeleton robot in the target cycle exceeds the preset fluctuation threshold, the human-machine interaction state is in the reverse force state, and the amplitude of the joint interaction power exceeds the corresponding force limit, it is determined that the target cycle is in an emergency braking state. The difference between the fluctuation characteristic value of the target cycle and the fluctuation threshold is calculated to obtain the stiffness adjustment amount of the target cycle.

[0010] Optionally, when the fluctuation characteristic value of the exoskeleton robot in the target cycle exceeds a preset fluctuation threshold, and the human-machine interaction state is in a reverse force state, and the amplitude of the joint interaction power exceeds the corresponding force limit, it is determined that the target cycle is in an emergency braking state. After calculating the difference between the fluctuation characteristic value of the target cycle and the fluctuation threshold to obtain the stiffness adjustment amount of the target cycle, the method further includes: When the stiffness adjustment amount of the target period is greater than zero, the stiffness adjustment coefficient of the target period is determined according to the ratio of the stiffness adjustment amount to the preset stiffness zero transition reference amount. Using the target gait stage of the target cycle as an index, a query is performed in a preset benchmark mapping table to extract the target support stiffness corresponding to the target gait stage; The output stiffness of the exoskeleton robot joint during the target period is obtained by multiplying the stiffness adjustment coefficient of the target period with the target support stiffness.

[0011] Secondly, a control system for an exoskeleton robot is provided, the system comprising: The first calculation module is used to calculate the joint interaction power of the target cycle based on the correlation between the joint angular velocity of the exoskeleton robot and the human-machine interaction torque in the target cycle; wherein, the human-machine interaction torque characterizes the interaction torque between the exoskeleton robot and the human body, and the joint interaction power is used to characterize the intensity of the human-machine interaction. The analysis module is used to analyze the amplitude distribution of the interaction power of multiple historical joints before the target cycle to obtain the fluctuation characteristic value of the exoskeleton robot in the target cycle; the fluctuation characteristic value characterizes the degree of force waveform distortion of the exoskeleton robot in the target cycle. The determination module is used to perform gait analysis on the user during the target cycle, obtain the target gait stage of the target cycle, and determine the upper limit of force corresponding to the target cycle based on the target gait stage. The second calculation module is used to determine that the target cycle is in an emergency braking state when the fluctuation characteristic value of the exoskeleton robot in the target cycle exceeds a preset fluctuation threshold, the human-machine interaction state in the target cycle is in a reverse force state, and the amplitude of the joint interaction power in the target cycle exceeds the corresponding force limit. Based on the fluctuation characteristic value of the target cycle, the module calculates the stiffness adjustment amount of the target cycle. The human-machine interaction state is determined according to the sign of the joint interaction power, and the stiffness adjustment amount is used to reduce the output stiffness of the exoskeleton robot joints.

[0012] Optionally, based on the correlation between the joint angular velocity and human-robot interaction torque of the exoskeleton robot in the target cycle, the joint interaction power of the target cycle is calculated, including: During the target cycle, the angular positions of the joints of the exoskeleton robot are collected, and the angular positions of adjacent cycles are differentially calculated to obtain the joint angular velocity of the target cycle. During the target cycle, the torque of human-machine interaction on the exoskeleton robot is collected by a torque sensor; The joint interaction power of the exoskeleton robot during the target cycle is obtained by multiplying the joint angular velocity and the human-machine interaction torque.

[0013] This application offers the following advantages: Joint-interactive power quantization quantizes the interaction intensity between the exoskeleton robot and the human body. Simultaneously, by capturing the distortion characteristics of the force waveform with extremely low computational overhead, it enables rapid identification of emergency braking actions, avoiding recognition delays caused by excessively long data accumulation windows. State determination based on three conditions effectively distinguishes between normal walking disturbances and genuine emergency braking, preventing false triggering caused by mechanical vibration or electrical noise interference. This allows the exoskeleton to quickly adapt to the human body's avoidance actions, reducing the additional physical exertion and discomfort experienced by the wearer during emergency stops. While ensuring safe emergency braking response, it enhances the practicality and human-machine interface of the exoskeleton robot. Attached Figure Description

[0014] Figure 1 A flowchart of a control method for an exoskeleton robot in one embodiment; Figure 2 This is a schematic diagram of the control system of an exoskeleton robot in one embodiment. Detailed Implementation

[0015] The specific scheme of the control method for an exoskeleton robot provided in this application is described below with reference to the accompanying drawings. For example... Figure 1 As shown, the method includes: S11. Calculate the joint interaction power of the exoskeleton robot in the target cycle based on the correlation between the joint angular velocity and the human-machine interaction torque in the target cycle.

[0016] Among them, the human-machine interaction torque characterizes the interaction torque between the exoskeleton robot and the human body, and the joint interaction power is used to characterize the intensity of the human-machine interaction.

[0017] This application is applied to emergency braking scenarios of exoskeleton robots in unstructured environments such as gravel roads or irregular steps. It can be implemented based on the microcontroller of the exoskeleton robot. The microcontroller continuously executes control loops at fixed time intervals, such as 1 millisecond per cycle, and the target cycle is any cycle.

[0018] In one embodiment, the joint interaction power for the target cycle is calculated based on the correlation between the joint angular velocity and human-robot interaction torque of the exoskeleton robot during the target cycle, including: During the target cycle, the angular positions of the joints of the exoskeleton robot are collected, and the angular positions of adjacent cycles are differentially calculated to obtain the joint angular velocity of the target cycle. During the target cycle, the torque of human-machine interaction on the exoskeleton robot is collected by a torque sensor; The joint interaction power of the exoskeleton robot during the target cycle is obtained by multiplying the joint angular velocity and the human-machine interaction torque.

[0019] During the target cycle, the angular positions of the exoskeleton robot's joints are acquired. These joints primarily refer to the lower limb joints, such as the hip, knee, or ankle joints. An absolute encoder can be installed at the motor output, and the angular position for the target cycle is obtained by reading the encoder values. The angular positions of adjacent cycles are then differentially calculated to obtain the joint angular velocity for the target cycle. This differential calculation method only requires storing the angular position data from the previous cycle, occupying minimal memory space and making it suitable for high-frequency execution in microcontrollers.

[0020] Simultaneously, torque sensors are used to collect the human-machine interaction torque of the exoskeleton robot. These sensors can be placed at the strap connection between the exoskeleton and the human body, or installed at the output shaft of the joint actuators. The torque sensors can directly measure the torque exerted by the human body on the exoskeleton and the reaction torque exerted by the exoskeleton on the human body. The joint interaction power for the target cycle is then calculated by multiplying the joint angular velocity of the exoskeleton robot in the target cycle by the human-machine interaction torque. The sign of the joint interaction power indicates the direction of energy flow between the human and the robot at the current moment. A positive value indicates that the direction of the human body's force is consistent with the direction of the exoskeleton's movement, and the human body is conforming to the auxiliary movement of the exoskeleton. In this case, the human and robot are in a normal cooperative assistance state, which is a positive force state. A negative value indicates that the direction of the human body's force is opposite to the direction of the exoskeleton's movement, and the human body is resisting the predetermined movement of the exoskeleton, i.e., in a reverse force state. The amplitude of the joint interaction power indicates the intensity of the human-machine interaction.

[0021] S12. Analyze the distribution of the amplitude of the interaction power of multiple historical joints before the target cycle to obtain the fluctuation characteristic value of the exoskeleton robot in the target cycle.

[0022] The fluctuation characteristic value characterizes the degree of distortion of the force waveform of the exoskeleton robot during the target cycle. The force waveform refers to the curve of the joint interaction power of the exoskeleton robot changing with time.

[0023] To capture the spike distortion characteristics of the force waveform during emergency braking, the amplitude distribution of the interaction power of multiple historical joints prior to the target cycle can be analyzed.

[0024] In one embodiment, the distribution of amplitudes of multiple historical joint interaction powers prior to the target cycle is analyzed to obtain the fluctuation characteristic values ​​of the exoskeleton robot during the target cycle, including: Based on the degree of deviation between the amplitudes of the interaction power of multiple historical joints before the target cycle, the fluctuation characteristic value of the exoskeleton robot in the target cycle is calculated.

[0025] When a person walks normally, the amplitude of joint interaction power changes relatively smoothly, with small differences in amplitude between adjacent cycles, showing an overall slow trend. Therefore, the deviation between historical joint interaction power amplitudes is small. However, during emergency braking, the instantaneous contraction of human muscles generates braking torque, causing the amplitude of historical joint interaction power to rise sharply in a very short time, resulting in a spike-shaped waveform distortion. At this time, the deviation between historical amplitudes increases significantly, enabling rapid identification of emergency braking.

[0026] Specifically, based on the degree of deviation between the amplitudes of the interaction power of multiple historical joints prior to the target cycle, the fluctuation characteristic values ​​of the exoskeleton robot during the target cycle are calculated, including: Calculate the square of the amplitude of the historical joint interaction power of the previous cycle of the target cycle to obtain the square value of the target cycle; The product of the amplitude of the joint interaction power in the target period and the amplitude of the historical joint interaction power in the period two years prior to the target period is calculated to obtain the product value of the target period. The difference between the square of the target period and its product is used to obtain the fluctuation characteristic value of the target period.

[0027] Calculate the square of the amplitude of the historical joint interaction power of the previous cycle of the target cycle to obtain the square value of the target cycle. Calculate the product of the amplitude of the joint interaction power of the target cycle and the amplitude of the historical joint interaction power of the cycle two cycles prior to the target cycle to obtain the product value of the target cycle. Calculate the difference between the square value of the target cycle and its product value to obtain the fluctuation characteristic value of the target cycle.

[0028] Under normal, stable walking conditions, the force waveform changes smoothly, and the square of the target period is nearly equal to the product of the target periods. Subtracting the square results in a fluctuation characteristic value approaching zero, thus automatically canceling out low-frequency normal force fluctuations. However, during emergency braking, the force waveform exhibits a steep, spike-like distortion, with the square of the target period drastically exceeding the product of the target periods. Subtracting the square results in a significant positive difference. Furthermore, due to the nonlinear amplification effect of the square term, the fluctuation characteristic value can highlight the severity of the abrupt change with quadratic sensitivity. The larger the fluctuation characteristic value, the more severe the force waveform distortion, and the higher the probability of emergency braking.

[0029] S13. Perform gait analysis on the user during the target cycle to obtain the target gait stage of the target cycle, and determine the upper limit of force corresponding to the target cycle based on the target gait stage.

[0030] Because the mechanical requirements of the exoskeleton vary significantly in different gait stages, it is necessary to perform gait analysis on the user during the target period to determine the current target gait stage.

[0031] In one embodiment, gait analysis of the user is performed during a target period to obtain the target gait stage of the target period, including: During the target cycle, gait phase values ​​are calculated based on user motion data; wherein, the gait phase values ​​are percentage values ​​used to characterize the phase of the gait cycle; The gait phase value is determined as the target gait phase in which the target cycle is located; wherein, the target gait phase is the phase percentage obtained by performing phase localization of human gait.

[0032] During the target cycle, gait phase values ​​are calculated based on user motion data, including but not limited to: lower limb joint angles, angular velocities, plantar pressure, and posture information output by the inertial measurement unit (IMU). After collecting user motion data, gait phase values ​​can be calculated using preset gait phase estimation algorithms (such as phase mapping based on joint angle curves, finite state machines based on plantar pressure switching, etc.). Gait phase values ​​are percentage values ​​used to characterize the phase of the gait cycle, typically represented as percentages between 0% and 100% for a complete gait cycle. 0% corresponds to the heel strike initiation, 50% to the mid-stance, and 100% to the next heel strike on the same side. These gait phase values ​​are then used to determine the target gait phase of the target cycle.

[0033] The mechanical characteristics of different gait phases differ significantly. For example, the heel strike cushioning phase allows for a larger reverse force, while the mid-support phase does not allow for a large reverse force. Therefore, after obtaining the target gait phase, it is necessary to determine whether the current reverse force amplitude exceeds the normal allowable range of the target gait phase.

[0034] In one embodiment, determining the upper limit of force corresponding to the target cycle based on the target gait stage includes: Using the target gait stage of the target cycle as an index, a query is performed in a preset reference mapping table to extract the upper limit of force corresponding to the target gait stage, thus obtaining the upper limit of force corresponding to the target cycle; the upper limit of force corresponding to each stage in the human gait cycle is stored in the reference mapping table in advance.

[0035] The reference mapping table is pre-stored in the microcontroller's read-only memory. One method for constructing this reference mapping table is as follows: the complete gait cycle from 0% to 100% is equally divided into multiple discrete intervals, for example, each 5% interval is a discrete interval, totaling twenty discrete intervals. For each discrete interval, the reference mapping table has a pre-calibrated upper force limit. The upper force limit represents the maximum reverse human-machine interaction power amplitude allowed for normal walking of the exoskeleton within the corresponding gait interval. When the actual reverse human-machine interaction power amplitude exceeds the upper force limit, it indicates that the current force has exceeded the normal buffer range, and an abnormal situation may exist.

[0036] The calibration method for the upper limit of force can be obtained through experiments: In a laboratory environment, subjects wear exoskeleton robots and walk with a normal gait. The amplitude data of human-computer interaction power at each gait stage are collected. The higher quantile (such as the 95th percentile) of the data distribution at each stage is taken as the upper limit of force for the corresponding stage. The experimental calibration method can fully reflect the differences in mechanical characteristics at different gait stages, making the setting of the upper limit of force more in line with the actual human movement law.

[0037] S14. When the fluctuation characteristic value of the exoskeleton robot in the target cycle exceeds the preset fluctuation threshold, the human-machine interaction state in the target cycle is in the reverse force state, and the amplitude of the joint interaction power in the target cycle exceeds the corresponding force limit, it is determined that the target cycle is in an emergency braking state, and the stiffness adjustment amount of the target cycle is calculated based on the fluctuation characteristic value of the target cycle.

[0038] The human-machine interaction state is determined based on the sign of the joint interaction power, and the stiffness adjustment is used to reduce the output stiffness of the exoskeleton robot joints.

[0039] In one embodiment, when the fluctuation characteristic value of the exoskeleton robot in the target cycle exceeds a preset fluctuation threshold, the human-machine interaction state in the target cycle is in a reverse force state, and the amplitude of the joint interaction power in the target cycle exceeds the corresponding force limit, it is determined that the target cycle is in an emergency braking state. The stiffness adjustment amount of the target cycle is calculated based on the fluctuation characteristic value of the target cycle, including: When the joint interaction power of the target cycle is negative, the human-machine interaction state of the target cycle is determined to be the reverse force state. When the fluctuation characteristic value of the exoskeleton robot in the target cycle exceeds the preset fluctuation threshold, the human-machine interaction state is in the reverse force state, and the amplitude of the joint interaction power exceeds the corresponding force limit, it is determined that the target cycle is in an emergency braking state. The difference between the fluctuation characteristic value of the target cycle and the fluctuation threshold is calculated to obtain the stiffness adjustment amount of the target cycle.

[0040] When the joint interaction power of the target cycle is negative, the human-machine interaction state of the target cycle is determined to be a reverse force state. As mentioned earlier, the sign of the joint interaction power is determined by the direction of angular velocity and torque. When the sign of the joint interaction power is negative, it indicates that the direction of the human body's force is opposite to the predetermined direction of the exoskeleton's movement, and the human-machine interaction state is a reverse force state. That is, the human body is resisting the movement of the exoskeleton and may be in an emergency braking state. It should be noted that this reverse force state is only a necessary but not sufficient condition for emergency braking, because a reverse force may also briefly occur during the normal heel strike cushioning process.

[0041] When the fluctuation characteristic value of the exoskeleton robot in the target cycle exceeds the preset fluctuation threshold, the human-machine interaction state is in a reverse force state, and the amplitude of the joint interaction power exceeds the corresponding force limit, the target cycle is determined to be in an emergency braking state. A fluctuation characteristic value exceeding the fluctuation threshold indicates a severe distortion of the force waveform; a reverse force state indicates that the human body is resisting the exoskeleton's movement; and the amplitude of the joint interaction power exceeding the force limit indicates that the intensity of the reverse force exceeds the normal range of the current gait phase. Only when all three conditions are met simultaneously is a genuine emergency braking action confirmed.

[0042] If the target cycle no longer meets the above three conditions simultaneously, but the system was already in an emergency braking state in the previous cycle, the following judgment rules are executed: when the human-machine interaction state of the target cycle is still in a reverse force state and the amplitude of the joint interaction power still exceeds the corresponding upper limit of force, it is determined that the system maintains the emergency braking state, and the stiffness adjustment amount of the target cycle is set to be equal to the stiffness adjustment amount of the previous cycle; when the human-machine interaction state of the target cycle is no longer in a reverse force state, or the amplitude of the joint interaction power is not greater than the corresponding upper limit of force, it is determined that the high-load resistance period of emergency braking has ended, and the stiffness adjustment amount of the target cycle is set to zero.

[0043] After confirming an emergency braking state, the difference between the fluctuation characteristic value of the target period and the fluctuation threshold is calculated to obtain the stiffness adjustment amount of the target period. The larger the stiffness adjustment amount, the more intense the emergency braking, and the greater the stiffness adjustment required. The fluctuation threshold is calibrated as follows: the high-frequency fluctuation value of the power output of the operator is collected and recorded in advance in samples of normal walking on flat ground by test personnel. The 95th percentile value of this data distribution is taken as the fluctuation threshold. The fluctuation threshold is used to distinguish between the normal gradual change of force on the human body and the abnormal peak distortion during emergency stop.

[0044] The above judgment rules can prevent the equipment from being triggered falsely. If only the amplitude of the joint interaction power is used for judgment, normal large step landing buffer can easily cause false stop. If only the waveform is used for judgment, the high-frequency electrical noise generated by the slightest external metal collision can also easily cause false stop. If only the fluctuation characteristic value is used for judgment, the brief reverse force during the normal gait switching process can also cause false trigger. By judging through three conditions, it is ensured that stiffness adjustment is only activated when a real emergency braking occurs.

[0045] After determining the stiffness adjustment amount, this amount needs to be converted into actual control commands and sent to the servo drive. However, the stiffness adjustment amount is a value with specific physical dimensions, while the servo drive receives stiffness parameters with mechanical dimensions. The two dimensions do not match, so they cannot be used directly. At the same time, the stiffness adjustment also needs to consider the safety of the transition range. If the adjustment is too fast, the exoskeleton may completely lose its support, causing the wearer to lose balance. If the adjustment is too fast, it may cause mechanical impact, causing discomfort to the wearer.

[0046] Therefore, in one embodiment, when the fluctuation characteristic value of the exoskeleton robot in the target cycle exceeds a preset fluctuation threshold, and the human-machine interaction state is a reverse force state, and the amplitude of the joint interaction power exceeds the corresponding force limit, it is determined that the target cycle is in an emergency braking state. After calculating the difference between the fluctuation characteristic value of the target cycle and the fluctuation threshold to obtain the stiffness adjustment amount of the target cycle, the method further includes: When the stiffness adjustment amount of the target period is greater than zero, the stiffness adjustment coefficient of the target period is determined according to the ratio of the stiffness adjustment amount to the preset stiffness zero transition reference amount. Using the target gait stage of the target cycle as an index, a query is performed in a preset benchmark mapping table to extract the target support stiffness corresponding to the target gait stage; The output stiffness of the exoskeleton robot joint during the target period is obtained by multiplying the stiffness adjustment coefficient of the target period with the target support stiffness.

[0047] When the stiffness adjustment amount of the target period t When the value is greater than zero, the stiffness adjustment amount is compared with the preset stiffness zeroing transition reference amount. The ratio is used to determine the stiffness adjustment coefficient for the target period t. The calculation formula is: Among them, the stiffness-to-zero transition reference value. Derived from the factory calibration program, the system records the upper limit of the high-frequency power fluctuation value generated during passive emergency stop under rated maximum load, and obtains the stiffness zeroing transition reference value. It represents the percentage of the current stress distortion relative to the maximum allowable distortion limit of the system, 1 minus It represents the proportion of remaining support stiffness that the system should retain. The function is used to truncate negative numbers, ensuring that the stiffness adjustment coefficient is reduced to a minimum of 0. The more intense the braking, the closer the stiffness adjustment coefficient is to zero, and the greater the degree to which the exoskeleton abandons its original motion trajectory.

[0048] When the stiffness adjustment amount of the target period t is equal to zero, and the stiffness adjustment coefficient of the previous period is less than 1, it is determined that the emergency braking action has ended, but the support force of the exoskeleton has not yet fully recovered to the normal state. An incremental calculation is then performed to obtain the stiffness adjustment coefficient of the target period. The calculation formula is: The min() function is used to truncate values ​​greater than 1. The preset stiffness increment step can be set according to ergonomic and mechanical safety requirements, with a rated recovery time, for example, 500 milliseconds. The rated recovery time is the expected safe transition time for the exoskeleton to recover from a fully unloaded state to a normal support state. The stiffness increment step for each cycle is obtained by dividing the total number of cycles within the rated recovery time (500 milliseconds corresponds to 500 cycles when each cycle is 1 millisecond) by 1. The stiffness adjustment coefficient gradually accumulates in constant and small increments in each cycle, allowing the exoskeleton's support force to smoothly recover from a compliant state to a normal support state. This avoids the mechanical shock to the human body caused by instantaneous full load recovery and ensures complete recovery within a defined timeframe, guaranteeing the safety and comfort of human-machine collaboration.

[0049] If the stiffness adjustment amount of the target cycle is equal to zero and the stiffness adjustment coefficient of the previous cycle is equal to 1, it indicates that the system is in a completely normal walking state. Set the stiffness adjustment coefficient of the target cycle to 1 and do not perform any additional calculations.

[0050] Using the target gait stage of the target cycle as an index, a query is performed in a preset benchmark mapping table to extract the target support stiffness corresponding to the target gait stage. The benchmark mapping table is the same as the one used for the above-mentioned upper limit force query. The table also stores the target support stiffness corresponding to each gait stage. The target support stiffness represents the normal support parameters that the motor should maintain under the expected interference-free conditions. The target support stiffness can be calibrated according to the following steps: In a laboratory environment, a user wears the exoskeleton robot and walks on flat ground with a normal gait. The stiffness parameters required by the exoskeleton motor to provide normal auxiliary support are collected at each gait stage (such as the heel strike buffer period, mid-support period, and push-off period). The median or mean value of the data distribution for each stage is taken as the target support stiffness calibration value for the corresponding stage.

[0051] Then, based on the product of the stiffness adjustment coefficient and the target support stiffness of the target cycle, the output stiffness of the exoskeleton robot joint in the target cycle is obtained. This output stiffness is used as the impedance control stiffness parameter of the exoskeleton robot joint in the target cycle and is directly sent to the current loop or torque loop of the servo driver, so that the motor responds to the joint angle deviation according to the output stiffness. When emergency braking occurs, the stiffness adjustment coefficient is less than 1, the output stiffness is less than the target support stiffness, the motor output stiffness decreases, and the mechanical resistance provided when resisting the reverse pull of the human body is weakened. The user can drive the exoskeleton to complete the avoidance action with less additional physical effort.

[0052] This application resolves the contradiction between delay and false triggering in emergency braking recognition of exoskeleton robots through the above steps.

[0053] This application also provides a control system for an exoskeleton robot, such as... Figure 2 As shown, it includes: The first calculation module 21 is used to calculate the joint interaction power of the target cycle based on the correlation between the joint angular velocity of the exoskeleton robot and the human-machine interaction torque in the target cycle; wherein, the human-machine interaction torque characterizes the interaction torque between the exoskeleton robot and the human body, and the joint interaction power is used to characterize the intensity of the human-machine interaction. Analysis module 22 is used to analyze the amplitude distribution of the interaction power of multiple historical joints before the target cycle to obtain the fluctuation characteristic value of the exoskeleton robot in the target cycle; the fluctuation characteristic value characterizes the degree of force waveform distortion of the exoskeleton robot in the target cycle. The determination module 23 is used to perform gait analysis on the user during the target cycle, obtain the target gait stage of the target cycle, and determine the upper limit of force corresponding to the target cycle based on the target gait stage. The second calculation module 24 is used to determine that the target cycle is in an emergency braking state when the fluctuation characteristic value of the exoskeleton robot in the target cycle exceeds a preset fluctuation threshold, the human-machine interaction state of the target cycle is in a reverse force state, and the amplitude of the joint interaction power of the target cycle exceeds the corresponding force limit. Based on the fluctuation characteristic value of the target cycle, the module calculates the stiffness adjustment amount of the target cycle. The human-machine interaction state is determined according to the sign of the joint interaction power, and the stiffness adjustment amount is used to reduce the output stiffness of the exoskeleton robot joints.

[0054] Optionally, based on the correlation between the joint angular velocity and human-robot interaction torque of the exoskeleton robot in the target cycle, the joint interaction power of the target cycle is calculated, including: During the target cycle, the angular positions of the joints of the exoskeleton robot are collected, and the angular positions of adjacent cycles are differentially calculated to obtain the joint angular velocity of the target cycle. During the target cycle, the torque of human-machine interaction on the exoskeleton robot is collected by a torque sensor; The joint interaction power of the exoskeleton robot during the target cycle is obtained by multiplying the joint angular velocity and the human-machine interaction torque.

[0055] It should be noted that the system provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the system and method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0056] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A control method for an exoskeleton robot, characterized in that, The method includes: Based on the correlation between the joint angular velocity and human-machine interaction torque of the exoskeleton robot in the target cycle, the joint interaction power of the target cycle is calculated; whereby the human-machine interaction torque characterizes the interaction torque between the exoskeleton robot and the human body, and the joint interaction power is used to characterize the intensity of the human-machine interaction. The distribution of the amplitude of the interaction power of multiple historical joints before the target cycle is analyzed to obtain the fluctuation characteristic value of the exoskeleton robot in the target cycle; the fluctuation characteristic value characterizes the degree of force waveform distortion of the exoskeleton robot in the target cycle. Gait analysis is performed on the user during the target cycle to obtain the target gait stage of the target cycle, and the upper limit of force corresponding to the target cycle is determined based on the target gait stage. When the fluctuation characteristic value of the exoskeleton robot in the target cycle exceeds the preset fluctuation threshold, the human-machine interaction state in the target cycle is in the reverse force state, and the amplitude of the joint interaction power in the target cycle exceeds the corresponding force limit, it is determined that the target cycle is in an emergency braking state, and the stiffness adjustment amount of the target cycle is calculated based on the fluctuation characteristic value of the target cycle; wherein, the human-machine interaction state is determined according to the sign of the joint interaction power, and the stiffness adjustment amount is used to reduce the output stiffness of the exoskeleton robot joint.

2. The control method for an exoskeleton robot as described in claim 1, characterized in that, The step of calculating the joint interaction power for the target cycle based on the correlation between the joint angular velocity and human-machine interaction torque of the exoskeleton robot during the target cycle includes: During the target cycle, the angular positions of the joints of the exoskeleton robot are collected, and the angular positions of adjacent cycles are differentially calculated to obtain the joint angular velocity of the target cycle. During the target cycle, the torque of human-machine interaction on the exoskeleton robot is collected by a torque sensor; The joint interaction power of the exoskeleton robot during the target cycle is obtained by multiplying the joint angular velocity and the human-machine interaction torque.

3. The control method for an exoskeleton robot as described in claim 1, characterized in that, The analysis of the amplitude distribution of multiple historical joint interaction powers prior to the target cycle yields the fluctuation characteristic values ​​of the exoskeleton robot during the target cycle, including: Based on the degree of deviation between the amplitudes of the interaction power of multiple historical joints before the target cycle, the fluctuation characteristic value of the exoskeleton robot in the target cycle is calculated.

4. The control method for an exoskeleton robot as described in claim 3, characterized in that, The step of calculating the fluctuation characteristic value of the exoskeleton robot in the target period based on the degree of deviation between the amplitudes of the interaction power of multiple historical joints prior to the target period includes: Calculate the square of the amplitude of the historical joint interaction power of the previous cycle of the target cycle to obtain the square value of the target cycle; The product of the amplitude of the joint interaction power in the target period and the amplitude of the historical joint interaction power in the period two years prior to the target period is calculated to obtain the product value of the target period. The difference between the square of the target period and its product is used to obtain the fluctuation characteristic value of the target period.

5. The control method for an exoskeleton robot as described in claim 1, characterized in that, The step of performing gait analysis on the user during the target period to obtain the target gait stage of the target period includes: During the target cycle, gait phase values ​​are calculated based on user motion data; wherein, the gait phase values ​​are percentage values ​​used to characterize the phase of the gait cycle; The gait phase value is determined as the target gait phase in which the target cycle is located; wherein, the target gait phase is the phase percentage obtained by performing phase localization of human gait.

6. The control method for an exoskeleton robot as described in claim 5, characterized in that, The step of determining the upper limit of force corresponding to the target cycle based on the target gait stage includes: Using the target gait stage of the target cycle as an index, a query is performed in a preset reference mapping table to extract the upper limit of force corresponding to the target gait stage, thus obtaining the upper limit of force corresponding to the target cycle; the upper limit of force corresponding to each stage in the human gait cycle is stored in the reference mapping table in advance.

7. The control method for an exoskeleton robot as described in claim 1, characterized in that, When the fluctuation characteristic value of the exoskeleton robot in the target cycle exceeds a preset fluctuation threshold, the human-machine interaction state in the target cycle is in a reverse force state, and the amplitude of the joint interaction power in the target cycle exceeds the corresponding force limit, it is determined that the target cycle is in an emergency braking state. The stiffness adjustment amount of the target cycle is calculated based on the fluctuation characteristic value of the target cycle, including: When the joint interaction power of the target cycle is negative, the human-machine interaction state of the target cycle is determined to be the reverse force state. When the fluctuation characteristic value of the exoskeleton robot in the target cycle exceeds the preset fluctuation threshold, the human-machine interaction state is in the reverse force state, and the amplitude of the joint interaction power exceeds the corresponding force limit, it is determined that the target cycle is in an emergency braking state. The difference between the fluctuation characteristic value of the target cycle and the fluctuation threshold is calculated to obtain the stiffness adjustment amount of the target cycle.

8. The control method for an exoskeleton robot as described in claim 7, characterized in that, When the fluctuation characteristic value of the exoskeleton robot in the target cycle exceeds a preset fluctuation threshold, the human-machine interaction state is in a reverse force state, and the amplitude of the joint interaction power exceeds the corresponding force limit, the target cycle is determined to be in an emergency braking state. After calculating the difference between the fluctuation characteristic value of the target cycle and the fluctuation threshold to obtain the stiffness adjustment amount of the target cycle, the method further includes: When the stiffness adjustment amount of the target period is greater than zero, the stiffness adjustment coefficient of the target period is determined according to the ratio of the stiffness adjustment amount to the preset stiffness zero transition reference amount. Using the target gait stage of the target cycle as an index, a query is performed in a preset benchmark mapping table to extract the target support stiffness corresponding to the target gait stage; The output stiffness of the exoskeleton robot joint during the target period is obtained by multiplying the stiffness adjustment coefficient of the target period with the target support stiffness.

9. A control system for an exoskeleton robot, characterized in that, The system includes: The first calculation module is used to calculate the joint interaction power of the target cycle based on the correlation between the joint angular velocity of the exoskeleton robot and the human-machine interaction torque in the target cycle; wherein, the human-machine interaction torque characterizes the interaction torque between the exoskeleton robot and the human body, and the joint interaction power is used to characterize the intensity of the human-machine interaction. The analysis module is used to analyze the amplitude distribution of the interaction power of multiple historical joints before the target cycle to obtain the fluctuation characteristic value of the exoskeleton robot in the target cycle; the fluctuation characteristic value characterizes the degree of force waveform distortion of the exoskeleton robot in the target cycle. The determination module is used to perform gait analysis on the user during the target cycle, obtain the target gait stage of the target cycle, and determine the upper limit of force corresponding to the target cycle based on the target gait stage. The second calculation module is used to determine that the target cycle is in an emergency braking state when the fluctuation characteristic value of the exoskeleton robot in the target cycle exceeds a preset fluctuation threshold, the human-machine interaction state in the target cycle is in a reverse force state, and the amplitude of the joint interaction power in the target cycle exceeds the corresponding force limit. Based on the fluctuation characteristic value of the target cycle, the module calculates the stiffness adjustment amount of the target cycle. The human-machine interaction state is determined according to the sign of the joint interaction power, and the stiffness adjustment amount is used to reduce the output stiffness of the exoskeleton robot joints.

10. The control system for an exoskeleton robot as described in claim 9, characterized in that, Based on the correlation between the joint angular velocity and human-robot interaction torque of the exoskeleton robot during the target cycle, the joint interaction power during the target cycle is calculated, including: During the target cycle, the angular positions of the joints of the exoskeleton robot are collected, and the angular positions of adjacent cycles are differentially calculated to obtain the joint angular velocity of the target cycle. During the target cycle, the torque of human-machine interaction on the exoskeleton robot is collected by a torque sensor; The joint interaction power of the exoskeleton robot during the target cycle is obtained by multiplying the joint angular velocity and the human-machine interaction torque.

Images