Control method and device of flexible pneumatic joint unit and flexible exoskeleton system

Abstract

The invention discloses a control method and device of a flexible pneumatic joint unit and a flexible exoskeleton system.The control method comprises the following steps that the difference value of electromyographic signals of an antagonistic muscle group is calculated, and a corresponding torque intention is obtained; calculating the sum of the electromyographic signals of the antagonistic muscle group to obtain a corresponding stiffness intention; generating a corresponding expected moment based on the moment intention; generating a target pressure difference corresponding to each antagonistic pneumatic muscle pair based on the expected moment; generating a corresponding expected stiffness based on the stiffness intention; generating a basic pressure corresponding to each antagonistic pneumatic muscle pair based on the expected stiffness; based on the target pressure difference and the basic pressure, generating target pressure of each pneumatic muscle in the antagonistic pneumatic muscle pair; and adjusting the internal air pressure of the corresponding pneumatic muscle based on the target pressure. According to the method, the user is not forced to follow a preset track to reach the target position, but the physical compliance of the user is dynamically adjusted according to the natural force generation and tightening conditions of muscles during the action of the user.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a control method and device for a flexible pneumatic joint unit, and a flexible exoskeleton system comprising the control device. Background Art

[0002] The core challenge of exoskeleton robots is to achieve safe and natural human-machine interaction.

[0003] Pneumatic Soft Actuator (PSA) technology provides a revolutionary drive solution. Originating from the aerospace and military fields, it features lightweight, high power-to-weight ratio, and intrinsic safety.

[0004] Although flexible drive robots based on flexible pneumatic muscle technology and corresponding control methods have been disclosed, such as the invention patent with publication number CN113858201B, which proposes an adaptive variable impedance control method, system, and device for a flexible drive robot, this method performs position control based on the user's motion trajectory to predict the desired position. However, in actual use, it has disadvantages such as high cost, large computational complexity, and poor human-computer interaction experience. Summary of the Invention

[0005] The present invention addresses the shortcomings of high cost and large computational complexity in the prior art of position control based on predicting the desired position of the user's motion trajectory, and provides a control method and device for a flexible pneumatic joint unit and a flexible exoskeleton system.

[0006] In order to solve the above technical problems, the present invention is solved by the following technical solutions:

[0007] A control method for a flexible pneumatic joint unit, the flexible pneumatic joint unit comprising at least one antagonistic pneumatic muscle pair;

[0008] The joint where the flexible pneumatic joint unit is worn is used as a target joint, the antagonistic pneumatic muscle pair corresponds to the antagonistic muscle group at the target joint, and the flexible pneumatic joint unit is controlled based on the electromyographic signal of the antagonistic muscle group at the target joint, including the following steps:

[0009] Calculating the difference in electromyographic signals of the antagonistic muscle groups to obtain the corresponding torque intention;

[0010] Calculating the sum of the electromyographic signals of the antagonistic muscle groups to obtain the corresponding stiffness intention;

[0011] Based on the torque intention, generating a corresponding expected torque;

[0012] generating target pressure differences corresponding to each antagonistic pneumatic muscle pair based on the desired torque;

[0013] Based on the stiffness intention, generating a corresponding expected stiffness;

[0014] generating a base pressure corresponding to each antagonistic pneumatic muscle pair based on the desired stiffness;

[0015] generating a target pressure for each pneumatic muscle in the antagonistic pneumatic muscle pair based on the target pressure difference and the base pressure;

[0016] The internal air pressure of the corresponding pneumatic muscle is adjusted based on the target pressure.

[0017] As an implementable method, a gravity compensation step is further included, specifically:

[0018] Pre-configure the weight parameter and distance parameter of the limb segment corresponding to the user's target joint, wherein the weight parameter is the weight of the limb segment, and the distance parameter is the distance from the center of mass of the limb segment to the target joint;

[0019] Obtaining a corresponding gravitational torque based on the joint angle corresponding to the target joint, the weight parameter, and the distance parameter;

[0020] A corresponding target pressure difference is generated based on the gravitational moment and the desired moment.

[0021] As an implementable approach:

[0022] The antagonistic pneumatic muscle pair includes two pneumatic muscles in an antagonistic configuration;

[0023] Based on the base pressure and the air pressure difference corresponding to the antagonistic pneumatic muscle pair, a reference pressure of the pneumatic muscle corresponding to the agonist muscle is obtained by addition calculation, and a reference pressure of the pneumatic muscle corresponding to the antagonist muscle is obtained by subtraction calculation;

[0024] The obtained reference pressure is input into a corresponding hysteresis model, and the corresponding hysteresis torque is predicted by the hysteresis model;

[0025] generating a corresponding compensation pressure based on the hysteresis torque;

[0026] The reference pressure is hysteresis compensated based on the compensation pressure to obtain a corresponding target pressure, and the internal air pressure of the corresponding pneumatic muscle is adjusted based on the target pressure.

[0027] As an implementable embodiment, the specific steps of adjusting the internal air pressure of the corresponding pneumatic muscle based on the target pressure are:

[0028] Real-time monitoring of the internal air pressure corresponding to the pneumatic muscle to obtain the corresponding feedback air pressure;

[0029] Based on the feedback air pressure and the target pressure, the internal air pressure of the pneumatic muscle is adjusted by a PD-type PID with integral limitation until the feedback air pressure matches the target pressure.

[0030] As an implementable method, when the stiffness intention is obtained, a lock switching judgment step is also included;

[0031] When the stiffness intention exceeds a preset locking threshold, switching the locking state;

[0032] When the locking state is switched from locked to unlocked, a corresponding expected stiffness is generated based on the stiffness intention, and a corresponding base pressure of each antagonistic pneumatic muscle pair is generated based on the expected stiffness;

[0033] When the locking state is switched from unlocked to locked, the preset locking stiffness is used as the expected stiffness, and based on the expected stiffness, the corresponding basic pressure of each antagonistic pneumatic muscle pair is generated.

[0034] As an implementation method, a corresponding torque mapping function is pre-constructed for the user's target joint. Based on the torque mapping function, the corresponding expected torque is determined according to the obtained torque intention. The method for constructing the torque mapping function is as follows:

[0035] Wearing the flexible pneumatic joint unit on the user's target joint, and placing and fixing the joint angle of the target joint according to a preset reference angle;

[0036] Collecting the output torque corresponding to the extension or flexion of the target joint at different force levels, and simultaneously collecting the electromyographic signal difference corresponding to the antagonistic muscle group at the target joint;

[0037] Based on the least squares algorithm, the output torque and electromyographic signal difference at each force level are fitted to obtain the torque mapping function corresponding to the user's target joint.

[0038] As an implementable method, a corresponding stiffness mapping function is pre-constructed for the user's target joint to indicate the mapping relationship between the stiffness intention and the expected stiffness. The method for constructing the stiffness mapping function is:

[0039] Obtain the measured stiffness and electromyographic signals of the antagonist muscle groups corresponding to the target joint at different tension levels;

[0040] Based on the least squares algorithm, the measured stiffness and electromyographic signals at each tension level are fitted to obtain the stiffness mapping function corresponding to the user's target joint;

[0041] The measurement method of the stiffness and electromyographic signal of the antagonist muscle group corresponding to the target joint at the target tension level is as follows:

[0042] Wearing the flexible pneumatic joint unit on the target joint of the user, and placing and maintaining the joint angle of the target joint according to a preset stiffness test angle;

[0043] Make the antagonistic muscle groups corresponding to the target joints co-contract according to the target tension level;

[0044] Applying a corresponding thrust pulse to the target joint according to a preset disturbance torque through the flexible pneumatic joint unit, collecting the joint angle change of the target joint, and collecting the electromyographic signals corresponding to the antagonistic muscle group;

[0045] Based on the disturbance torque and the joint angle change, a corresponding measurement stiffness is obtained.

[0046] A control device for a flexible pneumatic joint unit, the flexible pneumatic joint unit comprising at least one antagonistic pneumatic muscle pair;

[0047] The joint on which the flexible pneumatic joint unit is worn is used as a target joint, the antagonistic pneumatic muscle pair corresponds to the antagonistic muscle group at the target joint, and the flexible pneumatic joint unit is controlled based on the electromyographic signal of the antagonistic muscle group at the target joint, including:

[0048] an intention decoupling module, configured to calculate the difference of the myoelectric signals of the antagonistic muscle groups to obtain a corresponding torque intention; further configured to calculate the sum of the myoelectric signals of the antagonistic muscle groups to obtain a corresponding stiffness intention; further configured to generate a corresponding expected torque based on the torque intention; further configured to generate a corresponding expected stiffness based on the stiffness intention;

[0049] a calculation module, configured to generate a target pressure differential corresponding to each antagonistic pneumatic muscle pair based on the desired torque, and further configured to generate a base pressure corresponding to each antagonistic pneumatic muscle pair based on the desired stiffness; and further configured to generate a target pressure for each pneumatic muscle in the antagonistic pneumatic muscle pair based on the target pressure differential and the base pressure;

[0050] The control module is configured to adjust the internal air pressure of the corresponding pneumatic muscle based on the target pressure.

[0051] As an implementable approach:

[0052] The intention decoupling module obtains the corresponding gravity torque based on the joint angle, weight parameter and distance parameter corresponding to the target joint, wherein the weight parameter is the weight of the limb segment corresponding to the user's target joint, and the distance parameter is the distance from the center of mass of the limb segment to the target joint;

[0053] The calculation module is configured to generate a corresponding target pressure difference based on the gravity torque and the expected torque.

[0054] A flexible exoskeleton system, comprising a control device, a drive device, flexible pneumatic joint units corresponding to joints to be assisted, and a detection unit;

[0055] The corresponding flexible pneumatic joint unit and the joint to be assisted corresponding to the detection unit are recorded as target joints;

[0056] The flexible pneumatic joint unit includes at least one antagonistic pneumatic muscle pair;

[0057] The detection unit comprises:

[0058] The myoelectric detection component is used to collect the myoelectric signals corresponding to the antagonistic muscle group corresponding to the target joint;

[0059] An air pressure detection component for collecting the internal air pressure corresponding to each pneumatic muscle in each antagonistic pneumatic muscle pair;

[0060] The control device is connected to each electromyographic detection component and each air pressure detection component, and is used to execute a control method for a flexible pneumatic joint unit as described in any one of claims 1 to 7, and adjust each first pneumatic muscle and the internal air pressure of each first pneumatic muscle through a driving device.

[0061] The present invention has significant technical effects due to the adoption of the above technical solutions:

[0062] Compared with the prior art solution of training an intention recognition model and predicting the target joint position through the intention recognition model, the present invention only requires simple calculations based on electromyographic signals, with low computing resource consumption and low requirements on hardware structure.

[0063] Compared with existing solutions that specify a target position or predict a target position based on a model and then drive the user to reach the target position, the present invention does not force the user to follow a predetermined trajectory to reach the specified position. Instead, it adjusts the physical stiffness of the joints in real time, allowing the exoskeleton to reduce its stiffness and become extremely compliant when flexible movement is required; and increase its stiffness to provide solid and reliable support when stable support is required.

[0064] The control method proposed in this invention allows users to intuitively control the exoskeleton's output torque and stiffness through the natural exertion and tension of their muscles, which is highly consistent with the body's own motion control mechanism. This makes human-computer interaction more natural and the user experience is improved.

[0065] The control method proposed in this invention is based on compliant control and uses flexible pneumatic muscles, which provides a double safety guarantee. Even in extreme situations, the system is essentially a soft and cushioned wearable device. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0067] Figure 1 is a schematic diagram of a process for controlling the internal air pressure of each pneumatic muscle in an antagonistic pneumatic muscle pair;

[0068] Figure 2 This is a schematic diagram of module connections for a control device for a flexible pneumatic joint according to the present invention;

[0069] Figure 3 A schematic diagram of module connections of a flexible exoskeleton system according to the present invention. DETAILED DESCRIPTION

[0070] The present invention will be further described in detail below with reference to the examples. The following examples are intended to explain the present invention but the present invention is not limited to the following examples.

[0071] The flexible pneumatic joint unit described in the present application will be worn to the target joint of the user in actual use, and the at least one antagonistic pneumatic muscle pair is two pneumatic muscles in an antagonistic configuration, and the antagonistic pneumatic muscles correspond to the antagonistic muscle group at the target joint;

[0072] Here, the target joint is an active joint (such as a hip joint or a knee joint), and the antagonistic muscle groups are two groups of muscles that antagonize each other at the target joint, namely, an extensor group and a flexor group, such as the quadriceps femoris and the hamstrings; the pneumatic muscles correspond to the muscle groups, that is, the antagonistic pneumatic muscle pair includes a pneumatic muscle corresponding to the extensor group and a pneumatic muscle corresponding to the flexor group;

[0073] Pneumatic Soft Actuator (PSA) refers to a type of actuator driven by compressed gas (usually air) and made of flexible materials (such as silicone, rubber, fabric, etc.). Its axial elongation and output force are positively correlated with the internal air pressure, and its structural stiffness also increases with the increase of internal air pressure.

[0074] This application uses pneumatic muscles (PSA) to simulate the collaborative working mode of flexor and extensor muscles in biology to assist the target joint;

[0075] The present application is an innovation in the control method of a flexible pneumatic joint unit, and does not limit the flexible pneumatic joint unit itself. It can control any existing disclosed flexible pneumatic joint unit having the antagonistic pneumatic muscle pair.

[0076] Reference Figure 1 This application specification provides an implementable method for controlling a flexible pneumatic joint unit, wherein controlling the flexible pneumatic joint unit based on the electromyographic signal of the antagonistic muscle group at the target joint comprises the following steps:

[0077] S100, Intent Decoupling:

[0078] The electromyographic signals corresponding to the antagonistic muscle groups at the target joint are decomposed into desired torque and desired stiffness.

[0079] S110, calculating the difference of the electromyographic signals of the antagonistic muscle groups to obtain the corresponding torque intention;

[0080] Reference Figure 1 , Figure 1 Medium green indicates that the current movement of the relevant parameters corresponds to the agonist muscle of the antagonist muscle group, and red indicates that the relevant parameters correspond to the antagonist muscle of the antagonist muscle group;

[0081] In this embodiment, the torque is intended to be EMG agonist -EMG antagonist , where EMG agonis is the amplitude of the electromyographic signal corresponding to the agonist muscle in the antagonist muscle group, EMG antagonist is the amplitude of the electromyographic signal corresponding to the antagonist muscle. agonist Significantly greater than EMG antagonist , it indicates that the user intends to generate a buckling moment.

[0082] When the flexor EMG is significantly greater than the extensor EMG, it indicates that the user intends to generate a flexion moment.

[0083] In actual application, two sets of myoelectric sensors are set at the antagonistic muscle groups to detect the user's force intention and muscle co-contraction level in real time;

[0084] During actual movement, the muscle groups corresponding to the agonist and antagonist muscles will change with the flexion and extension of the joints. In this embodiment, the judgment is made based on the collected electromyographic signals, that is, the electromyographic signals of the two groups of muscle groups in the antagonist muscle group are compared, and the muscle group with the larger amplitude is used as the agonist muscle, and the muscle group with the smaller amplitude is used as the antagonist muscle.

[0085] As an implementable method, EMG agonist and EMG antagonist is the value of the normalized electromyographic signal;

[0086] After normalizing the original electromyographic signal collected by the electromyographic sensor, the electromyographic signal used for decoupling in step S100 is obtained. The specific steps of normalization in this embodiment are:

[0087] Collect the EMG peak value of each muscle group in the user's antagonistic muscle group at the maximum force state in advance, recorded as EMG MVC ;

[0088] EMG based on muscle groups MVC The original electromyographic signal EMG of the muscle group raw Perform standard normalization processing, EMG norm =EMG raw / EMG MVC , get the relative activation EMG between 0 and 1 norm , that is, the corresponding electromyographic signal.

[0089] When the user uses it for the first time or is calibrating, guide the user to contract the corresponding muscle groups with maximum force for 2-3 seconds in turn, and record the EMG peak value of the corresponding muscle group at this time.

[0090] After the raw electromyographic signals are normalized, whether it is a strong man or a patient in the early stages of recovery, their "50% force level" will be expressed as a value close to 0.5 to eliminate the impact of individual physiological differences.

[0091] S120, calculating the sum of the electromyographic signals of the antagonistic muscle groups to obtain the corresponding stiffness intention;

[0092] In this embodiment, the stiffness is intended to be EMG agonist +EMG antagonist , where EMG agonis is the amplitude of the electromyographic signal corresponding to the agonist muscle in the antagonist muscle group, EMG antagonist is the amplitude of the electromyographic signal corresponding to the antagonist muscle in the antagonist muscle group. agonist and EMG antagonist When they increase simultaneously (i.e., muscles co-contract), it indicates that the user wants the joint to become more stable;

[0093] In practical applications, those skilled in the art may set weighted weights for the agonist muscles and the antagonist muscles to perform weighted summation, which is not limited in detail in this embodiment.

[0094] S130: Generate a corresponding expected torque based on the torque intention;

[0095] Based on the isometric measurement method, the relationship between the electromyographic signal difference corresponding to the antagonistic muscle group at the user's target joint and the user's output torque is measured in advance, thereby constructing a corresponding torque mapping function for the user's target joint;

[0096] Based on the torque mapping function, the corresponding expected torque can be determined according to the obtained torque intention;

[0097] The method for constructing the torque mapping function is:

[0098] The flexible pneumatic joint unit is worn on the target joint of the user, and the joint angle of the target joint is placed and fixed according to a preset reference angle. The reference angle of the target joint is determined based on biomechanics. For example, when the target joint is a knee joint, the reference angle is 90°.

[0099] The output torque corresponding to the extension or flexion of the target joint at different force levels is collected, and the myoelectric signal difference corresponding to the antagonistic muscle group at the target joint is collected (the collected original myoelectric signal is normalized to obtain the corresponding myoelectric signal, and the absolute value of the difference in the myoelectric signal amplitude corresponding to the two muscle groups in the antagonistic muscle group is calculated);

[0100] Based on the least squares algorithm, the output torque and electromyographic signal difference at each force level are fitted to obtain the torque mapping function corresponding to the user's target joint.

[0101] In practical applications, the obtained torque intention is input into the constructed torque mapping function, and the torque output by the torque mapping function is used as the corresponding expected torque.

[0102] As an implementable method, the output of the torque mapping function is optimized by the torque gain to obtain the corresponding desired torque;

[0103] A one-time baseline calibration is insufficient to cope with changes in user status (such as rehabilitation progress), while frequent calibration affects the user experience. To address this issue, we propose increasing or decreasing the torque gain based on user feedback, ensuring that the expected torque remains consistent with the user's status between baseline calibrations.

[0104] The torque output by the torque mapping function is weighted by the torque gain to obtain the corresponding expected torque.

[0105] Those skilled in the art may set the torque gain adjustment rules according to actual needs. In this embodiment, a coarse adjustment is first performed based on the power assist deviation degree reported by the user, and then a fine adjustment is performed based on the power assist deviation reported by the user until the user determines that the adjustment is successful. Specifically,

[0106] Preset the mapping data of the power assist deviation degree and the torque gain adjustment amount. According to the power assist deviation degree (too large, too large, appropriate, too small, too small) fed back by the user, obtain the torque gain adjustment amount and make a rough adjustment.

[0107] After the adjustment is completed, the power assistance deviation (too large, appropriate, too small) is gradually adjusted according to the unit adjustment value based on the user feedback until the user feedback shows that the power assistance is appropriate.

[0108] S140, generating a corresponding expected stiffness based on the stiffness intention;

[0109] A corresponding stiffness mapping function is constructed in advance for the user's target joint to indicate the mapping relationship between the stiffness intention and the expected stiffness. The method for constructing the stiffness mapping function is:

[0110] Obtain the measured stiffness and electromyographic signals of the antagonist muscle groups corresponding to the target joint at different tension levels;

[0111] Based on the least squares algorithm, the measured stiffness and electromyographic signals at each tension level are fitted to obtain the stiffness mapping function corresponding to the user's target joint;

[0112] The measurement method of the stiffness and electromyographic signal of the antagonist muscle group corresponding to the target joint at the target tension level is as follows:

[0113] The flexible pneumatic joint unit is worn on the target joint of the user, and the joint angle of the target joint is placed and maintained according to a preset stiffness test angle. Those skilled in the art can determine the stiffness test angle based on biomechanics;

[0114] Make the antagonistic muscle groups corresponding to the target joints co-contract according to the target tension level;

[0115] Through the flexible pneumatic joint unit, a corresponding thrust pulse is applied to the target joint according to a preset disturbance torque (which can be determined by those skilled in the art according to actual conditions), and the joint angle change Δθ of the target joint is collected. At the same time, the sum of the myoelectric signals corresponding to the antagonistic muscle groups is collected, that is, the collected original myoelectric signals are normalized to obtain the corresponding myoelectric signals, and the sum of the myoelectric signals corresponding to the two antagonistic muscle groups is calculated;

[0116] Based on the disturbance torque and the change in the joint angle, the corresponding measurement stiffness is obtained, and the measurement stiffness Kactual=τ / Δθ, where τ is the disturbance torque.

[0117] In practical applications, the stiffness mapping function is used to calculate the stiffness based on the stiffness intent and convert it into the expected stiffness.

[0118] As an implementable method, the result output by the stiffness mapping function is optimized by the stiffness gain to obtain the corresponding desired stiffness;

[0119] As mentioned above, a one-time baseline calibration is not enough to cope with changes in user status (such as rehabilitation progress), and too frequent calibration affects the user experience. To address this problem, it is proposed to increase or decrease the stiffness gain based on user feedback.

[0120] The stiffness output by the stiffness mapping function is weighted by the stiffness gain to obtain the corresponding expected stiffness.

[0121] Those skilled in the art can set the adjustment rules of the stiffness gain according to actual needs. In this embodiment, a rough adjustment is first performed based on the joint stability feedback from the user, and then a fine adjustment is performed based on the joint stability feedback from the user (too hard, relatively hard, appropriate, relatively soft, too soft) until the user determines that the adjustment is successful. The specific implementation method is consistent with the above-mentioned torque gain, so it will not be repeated here.

[0122] S200, pressure calculation:

[0123] For each antagonistic pneumatic muscle pair, the pressure calculation is performed according to the following steps to determine the target pressure of each pneumatic muscle in the antagonistic pneumatic muscle pair:

[0124] S210, calculating the target pressure difference and base pressure corresponding to the antagonistic pneumatic muscle pair;

[0125] generating a target pressure difference corresponding to the antagonistic pneumatic muscle pair based on the desired torque;

[0126] generating a base pressure corresponding to each antagonistic pneumatic muscle pair based on the desired stiffness;

[0127] The calculation formula is as follows:

[0128]

[0129] in:

[0130] Δp is the target pressure difference between the antagonistic pneumatic muscle pairs;

[0131] p co The basic pressure of the antagonistic pneumatic muscle pair, i.e., the preload pressure;

[0132] τ des is the corresponding expected torque.

[0133] K des is the corresponding expected stiffness.

[0134] f(·) is the mapping function from torque to pressure difference;

[0135] g(·) is the mapping function from stiffness to pressure;

[0136] The above mapping functions f(·) and g(·) can be derived from the physical model of the pneumatic muscle used, or obtained through experimental calibration. This embodiment adopts the method of physical model derivation;

[0137] The physical model used to derive the mapping function in this embodiment includes:

[0138] Yeoh's Model for Strain Energy, used to describe the hyperelastic mechanical behavior of the flexible material (silicone) inside the pneumatic muscle;

[0139] The formula is as follows:

[0140] w c =C 10 (I C -3)+C 20 (I C -3) 2

[0141] w c Used to describe the strain energy stored per unit volume of flexible materials (silicone);

[0142] C 10 and C 10 It is the characteristic parameter of silicone material, which is determined by material tensile test;

[0143] I C is the first Cauchy-Green strain invariant, a scalar that comprehensively describes the degree of stretching or compression of a material in all directions.

[0144] Pressure-Deformation Model, which establishes the relationship between the input air pressure p and the geometric deformation of the pneumatic muscle (represented by the opening angle θ);

[0145] The formula is as follows:

[0146]

[0147] in:

[0148] p is the driving air pressure applied inside the pneumatic muscle;

[0149] δ represents a small change in a variable (virtual displacement or virtual change in the principle of virtual work);

[0150] θ represents the opening angle of the external braided mesh sleeve of the pneumatic muscle;

[0151] r1, L1, and t1 are the inner radius, length, and wall thickness of the actuator in the inflated state, respectively;

[0152] r0, L0, and t0 are the inner radius, length, and tube wall thickness of the actuator in the initial uninflated state, respectively.

[0153] The force-pressure model is used to describe the axial output force F that a single PSA can generate under a specific input pressure p.

[0154] The formula is as follows:

[0155]

[0156] in:

[0157] F is the output force generated by the pneumatic muscle in the axial direction (length direction);

[0158] r1′ represents the rate of change (derivative) of the pneumatic muscle radius r1 to its own length L1, which describes the radial contraction characteristics of the actuator during the extension process;

[0159] Wc represents the total strain energy stored.

[0160] As an implementable method, it further includes the step of gravity compensation;

[0161] Human limbs have their own weight. At different angles, the torque required to overcome gravity is different. If only the user's electromyographic signals are used for control, the user will need to exert extra effort to counteract the weight of the limb itself.

[0162] This embodiment makes the user feel that his limbs are "weightless" through gravity compensation, so that during use, the collected electromyographic signals more purely reflect his active movement intention rather than the compensatory intention.

[0163] The specific steps include:

[0164] Pre-configure the weight parameter m and distance parameter l of the limb segment corresponding to the user's target joint, where the weight parameter m is the weight of the limb segment and the distance parameter l is the distance from the center of mass of the limb segment to the target joint;

[0165] Based on the joint angle θ corresponding to the target joint, the weight parameter m and the distance parameter l, the corresponding gravity moment τ is obtained g , τ g =m·g·l·cos(θ), where g represents the acceleration due to gravity.

[0166] Based on the gravitational moment τ g and the desired torque τ des Generate the corresponding target pressure difference, that is, Δp = f(τ des +τ g ), the mapping function f(·) from torque to pressure difference is based on the gravity torque τ g and the desired torque τ des The sum of determines the pressure difference.

[0167] S220, generating a target pressure for each pneumatic muscle in the antagonistic pneumatic muscle pair based on the target pressure difference and the base pressure;

[0168] The antagonistic pneumatic muscle pair includes two pneumatic muscles in an antagonistic configuration, and the steps of generating the target pressure of the pneumatic muscles are:

[0169] S221. Based on the base pressure and the air pressure difference corresponding to the antagonistic pneumatic muscle pair, obtain a reference pressure of the pneumatic muscle corresponding to the agonist muscle by an addition operation, and obtain a reference pressure of the pneumatic muscle corresponding to the antagonist muscle by a subtraction operation;

[0170] The calculation formula is:

[0171]

[0172] in:

[0173] p target_agonist Indicates the reference pressure of the pneumatic muscle corresponding to the agonist muscle;

[0174] p target_antagonist Indicates the reference pressure of the pneumatic muscle corresponding to the antagonist muscle;

[0175] P co Indicates base pressure;

[0176] Δp represents the target pressure difference.

[0177] S222, inputting the obtained reference pressure into a corresponding hysteresis model, and using the hysteresis model to predict the corresponding hysteresis torque;

[0178] During factory delivery or regular maintenance, each pneumatic muscle needs to be calibrated for its dynamic characteristics to obtain time series data including input pressure, output force, and elongation. A corresponding hysteresis model can be constructed based on the time series data corresponding to each pneumatic muscle. For example, a corresponding hysteresis model can be constructed based on the existing publicly available improved Bouc-Wen model.

[0179] The input of the trained hysteresis model is the reference pressure corresponding to the current gas muscle, and the output is the predicted hysteresis torque.

[0180] Note: Those skilled in the art can train the hysteresis model of the corresponding pneumatic muscle based on the existing hysteresis model training method. This is not an innovation of this application and will not be introduced in detail in the specification.

[0181] S223, generating a corresponding compensation pressure based on the hysteresis torque;

[0182] Those skilled in the art can determine the feedforward gain K according to actual conditions. ff , based on the feedforward gain K ff Hysteresis torque τ hys Weighted, the corresponding compensation pressure p is obtainedcomp , p comp =K ff ·τ hys .

[0183] S224 : Perform hysteresis compensation on the reference pressure based on the compensation pressure to obtain a corresponding target pressure.

[0184] The target pressure is the sum of the compensation pressure and the reference pressure.

[0185] S400, pressure control:

[0186] The internal air pressure of the corresponding pneumatic muscle is adjusted based on the target pressure, and the specific steps are as follows:

[0187] Real-time monitoring of the internal air pressure corresponding to the pneumatic muscle to obtain the corresponding feedback air pressure;

[0188] Based on the feedback air pressure and the target pressure, the internal air pressure of the pneumatic muscle is adjusted by a PD-type PID with integral limitation until the feedback air pressure matches the target pressure.

[0189] As an implementable method, before calculating the expected stiffness based on the stiffness intention, a lock switching judgment step is further included;

[0190] When the stiffness intention exceeds a preset locking threshold, switching the locking state;

[0191] When the locking state is switched from locked to unlocked, based on the stiffness intention, a corresponding expected stiffness is generated according to the above step S140, and based on the expected stiffness, a corresponding base pressure of each antagonistic pneumatic muscle pair is generated;

[0192] When the locking state switches from unlocked to locked, the preset locking stiffness is directly used as the expected stiffness, and based on the expected stiffness, the corresponding basic pressure of each antagonistic pneumatic muscle pair is generated. At this time, the current joint angle will be recorded, and the target joint will be locked by increasing the basic pressure, allowing the user to easily maintain the current joint angle.

[0193] This embodiment triggers locking or unlocking by having the user tighten their muscles to increase the co-contraction state of the muscles. In an assisting scenario, natural assisting or maintaining the movement can be achieved according to the needs of the user.

[0194] like Figure 2 As shown, the present application specification provides an implementable embodiment of a control device for a flexible pneumatic joint unit, which controls the flexible pneumatic joint unit based on the electromyographic signal of the antagonistic muscle group at the target joint, including:

[0195] The intention decoupling module 110 is configured to calculate the difference of the electromyographic signals of the antagonistic muscle groups to obtain a corresponding torque intention; further configured to calculate the sum of the electromyographic signals of the antagonistic muscle groups to obtain a corresponding stiffness intention; further configured to generate a corresponding expected torque based on the torque intention; further configured to generate a corresponding expected stiffness based on the stiffness intention;

[0196] a calculation module 120 for generating a target pressure difference corresponding to each antagonistic pneumatic muscle pair based on the desired torque, and for generating a base pressure corresponding to each antagonistic pneumatic muscle pair based on the desired stiffness; and for generating a target pressure for each pneumatic muscle in the antagonistic pneumatic muscle pair based on the target pressure difference and the base pressure;

[0197] The control module 130 is configured to adjust the internal air pressure of the corresponding pneumatic muscle based on the target pressure.

[0198] In this embodiment, the control module 130 adopts a pressure closed-loop controller.

[0199] As an implementable approach:

[0200] The intention decoupling module 110 obtains the corresponding gravity torque based on the joint angle, weight parameter, and distance parameter corresponding to the target joint, wherein the weight parameter is the weight of the limb segment corresponding to the user's target joint, and the distance parameter is the distance from the center of mass of the limb segment to the target joint;

[0201] The calculation module 120 is configured to generate a corresponding target pressure difference based on the gravity torque and the expected torque.

[0202] like Figure 3 As shown, this application specification provides an implementation method of a control device for a flexible pneumatic joint unit, a flexible exoskeleton system, including a control device 100, a drive device 200, flexible pneumatic joint units 300 corresponding to each joint to be assisted, and a detection unit 400. For the sake of neatness of the drawings, Figure 3 The connection between the detection unit 400 and the control device 100 is omitted;

[0203] The corresponding flexible pneumatic joint unit 300 and the joint to be assisted corresponding to the detection unit 400 are recorded as target joints;

[0204] The flexible pneumatic joint unit 300 includes at least one antagonistic pneumatic muscle pair;

[0205] The detection unit 400 includes:

[0206] The myoelectric detection component is used to collect the myoelectric signals corresponding to the antagonistic muscle group corresponding to the target joint;

[0207] An air pressure detection component for collecting the internal air pressure corresponding to each pneumatic muscle in each antagonistic pneumatic muscle pair;

[0208] The control device 100 is connected to each electromyographic detection component and each air pressure detection component, and is used to execute any one of the control methods of a flexible pneumatic joint unit 300 described above, and adjust each first pneumatic muscle and the internal air pressure of each first pneumatic muscle through the driving device 200.

[0209] As an implementable embodiment, the detection unit 400 further includes a joint angle encoder capable of monitoring the joint angle of the target joint in real time.

[0210] Case:

[0211] The flexible exoskeleton system disclosed in the above embodiment is used to assist hemiplegic patients in sit-to-stand and walking assistance training;

[0212] Phase 1: Initialization and wearing;

[0213] The flexible exoskeleton system is turned on, and the control device 100 performs a self-check and performs baseline noise calibration on all EMG channels. The air pressure in all pneumatic muscle cavities is set to a preset initial value (such as p co =0.05MPa), at this time, each flexible pneumatic joint unit 300 is very soft and compliant, which is convenient for the patient to wear and adjust the posture.

[0214] Wear the flexible pneumatic joint units 300 corresponding to the hip and knee joints of the flexible exoskeleton to the corresponding joints;

[0215] The electromyography detection component (EMG surface electrode) was attached to the quadriceps femoris (extensor muscles) and hamstring muscles (flexor muscles) of both legs.

[0216] Phase 2: Variable Impedance Control in Sit-to-Stand Tasks:

[0217] When a patient prepares to stand up from a chair, they naturally tense their leg muscles. At this point, the EMG detection component detects a significant increase in the raw EMG signals of their quadriceps (extensors) and hamstrings (flexors).

[0218] The control device 100 pre-processes the original electromyographic signals of the quadriceps and hamstring muscles to obtain standard normalized electromyographic signals.

[0219] The control device 100 knows, based on the sum of the myoelectric signals of the quadriceps femoris and the hamstring muscles (ie, the stiffness intention), that the user has a strong intention to improve joint stability, that is, a high stiffness is required.

[0220] At the same time, since the amplitude of the electromyographic signal of the quadriceps femoris is much larger than that of the hamstrings, the control device 100 knows that the user needs a strong stretching torque to support the body based on the difference between the electromyographic signals of the quadriceps femoris and the hamstrings (i.e., torque intention).

[0221] The control device 100 calculates the pressure information of the corresponding antagonistic pneumatic unit pair based on the pre-built mapping formula according to the torque intention and stiffness intention, that is, the target pressure difference Δp (0.2 MPa) and the basic pressure p co (0.3MPa).

[0222] The agonist and antagonist muscles can be distinguished based on the electromyographic signals of the quadriceps and hamstrings. In this case, the quadriceps is the agonist and the hamstrings is the antagonist. In this case, the reference pressure of the pneumatic muscle is calculated based on the target pressure difference and the base pressure, and based on the state of the corresponding muscle group of the pneumatic muscle.

[0223] In this case, the reference pressure corresponding to the quadriceps femoris is 0.4 MPa, and the reference pressure corresponding to the antagonist muscle is 0.2 MPa.

[0224] The control device 100 performs corresponding hysteresis compensation on the determined reference pressure, determines the target pressure, and inflates the two pneumatic muscles based on the target pressure using PD-type PID control with integral limitation, so that the flexible pneumatic joint corresponding to the knee joint can provide powerful extension assistance, helping the patient to complete the standing movement smoothly and safely.

[0225] Phase 3: Dynamic compliance regulation during walking tasks:

[0226] Smooth control of the swing phase:

[0227] The patient starts walking, taking his healthy leg as an example. When the leg enters the swing phase, his muscles are in a relaxed state. At this time, the corresponding electromyographic signal amplitude is very low, so the sum of his electromyographic signal amplitudes is low, so that the control device 100 knows, based on the corresponding stiffness intention, that the user needs low stiffness to achieve flexible swinging.

[0228] The control device 100 generates a lower basic pressure based on the obtained stiffness intention, thereby reducing the stiffness of the antagonistic pneumatic muscles, and the resistance brought by the corresponding flexible pneumatic joints becomes extremely small, which hardly hinders the user's leg swinging movement, thereby achieving a high degree of motion compliance.

[0229] Stiffness control of the support phase:

[0230] As the leg enters the stance phase and prepares to bear weight, the muscles naturally contract to stabilize the joint.

[0231] The control device 100 detects that the stiffness is intended to be increased again, and determines that the user requires high stiffness.

[0232] The control device 100 generates a higher base pressure based on the obtained stiffness intention, so that the antagonistic pneumatic muscle pair is "tightened" and the corresponding flexible pneumatic joint exoskeleton joint becomes "hard", providing reliable support for the user's load-bearing legs and ensuring stability during walking.

[0233] From the above, it can be seen that the flexible exoskeleton system provided by the present invention no longer executes a fixed trajectory, but is able to sense the user's demand for force and stiffness in real time and dynamically change its own physical properties.

[0234] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0235] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0236] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, apparatus, 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.

[0237] The present invention is described with reference to the flowcharts and / or block diagrams of the method, terminal device (system), and computer program product according to the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, 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 terminal device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the process in the flowchart and / or block diagram. 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.

[0238] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1The function specified in one or more boxes.

[0239] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement 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.

Claims

1. A method for controlling a flexible pneumatic joint unit, wherein the flexible pneumatic joint unit comprises at least one antagonistic pneumatic muscle pair; It is characterized in that The joint where the flexible pneumatic joint unit is worn is used as a target joint, the antagonistic pneumatic muscle pair corresponds to the antagonistic muscle group at the target joint, and the flexible pneumatic joint unit is controlled based on the electromyographic signal of the antagonistic muscle group at the target joint, including the following steps: Calculating the difference in electromyographic signals of the antagonistic muscle groups to obtain the corresponding torque intention; Calculating the sum of the electromyographic signals of the antagonistic muscle groups to obtain the corresponding stiffness intention; Based on the torque intention, generating a corresponding expected torque; generating target pressure differences corresponding to each antagonistic pneumatic muscle pair based on the desired torque; Based on the stiffness intention, generating a corresponding expected stiffness; generating a base pressure corresponding to each antagonistic pneumatic muscle pair based on the desired stiffness; generating a target pressure for each pneumatic muscle in the antagonistic pneumatic muscle pair based on the target pressure difference and the base pressure; The internal air pressure of the corresponding pneumatic muscle is adjusted based on the target pressure.

2. The control method of a flexible pneumatic joint unit according to claim 1, characterized in that: It also includes gravity compensation steps, specifically: Pre-configure the weight parameter and distance parameter of the limb segment corresponding to the user's target joint, wherein the weight parameter is the weight of the limb segment, and the distance parameter is the distance from the center of mass of the limb segment to the target joint; Obtaining a corresponding gravitational torque based on the joint angle corresponding to the target joint, the weight parameter, and the distance parameter; A corresponding target pressure difference is generated based on the gravitational moment and the desired moment.

3. A control method for a flexible pneumatic joint unit according to claim 1 or 2, characterized in that: The antagonistic pneumatic muscle pair includes two pneumatic muscles in an antagonistic configuration; Based on the base pressure and the air pressure difference corresponding to the antagonistic pneumatic muscle pair, a reference pressure of the pneumatic muscle corresponding to the agonist muscle is obtained by addition calculation, and a reference pressure of the pneumatic muscle corresponding to the antagonist muscle is obtained by subtraction calculation; The obtained reference pressure is input into a corresponding hysteresis model, and the corresponding hysteresis torque is predicted by the hysteresis model; generating a corresponding compensation pressure based on the hysteresis torque; The reference pressure is hysteresis compensated based on the compensation pressure to obtain a corresponding target pressure, and the internal air pressure of the corresponding pneumatic muscle is adjusted based on the target pressure.

4. The control method of a flexible pneumatic joint unit according to claim 3, characterized in that: The specific steps of adjusting the internal air pressure of the corresponding pneumatic muscle based on the target pressure are: Real-time monitoring of the internal air pressure corresponding to the pneumatic muscle to obtain the corresponding feedback air pressure; Based on the feedback air pressure and the target pressure, the internal air pressure of the pneumatic muscle is adjusted by a PD-type PID with integral limitation until the feedback air pressure matches the target pressure.

5. A control method for a flexible pneumatic joint unit according to claim 1 or 2, characterized in that: When the stiffness intention is obtained, a lock switching judgment step is also included; When the stiffness intention exceeds a preset locking threshold, switching the locking state; When the locking state is switched from locked to unlocked, a corresponding expected stiffness is generated based on the stiffness intention, and a corresponding base pressure of each antagonistic pneumatic muscle pair is generated based on the expected stiffness; When the locking state is switched from unlocked to locked, the preset locking stiffness is used as the expected stiffness, and based on the expected stiffness, the corresponding basic pressure of each antagonistic pneumatic muscle pair is generated.

6. A control method for a flexible pneumatic joint unit according to claim 1 or 2, characterized in that: A corresponding torque mapping function is constructed in advance for the user's target joint. Based on the torque mapping function, the corresponding expected torque is determined according to the obtained torque intention. The method for constructing the torque mapping function is: Wearing the flexible pneumatic joint unit on the user's target joint, and placing and fixing the joint angle of the target joint according to a preset reference angle; Collecting the output torque corresponding to the extension or flexion of the target joint at different force levels, and simultaneously collecting the electromyographic signal difference corresponding to the antagonistic muscle group at the target joint; Based on the least squares algorithm, the output torque and electromyographic signal difference at each force level are fitted to obtain the torque mapping function corresponding to the user's target joint.

7. A control method for a flexible pneumatic joint unit according to claim 1 or 2, characterized in that: A corresponding stiffness mapping function is constructed in advance for the user's target joint to indicate the mapping relationship between the stiffness intention and the expected stiffness. The method for constructing the stiffness mapping function is: Obtain the measured stiffness and electromyographic signals of the antagonist muscle groups corresponding to the target joint at different tension levels; Based on the least squares algorithm, the measured stiffness and electromyographic signals at each tension level are fitted to obtain the stiffness mapping function corresponding to the user's target joint; The measurement method of the stiffness and electromyographic signal of the antagonist muscle group corresponding to the target joint at the target tension level is as follows: Wearing the flexible pneumatic joint unit on the target joint of the user, and placing and maintaining the joint angle of the target joint according to a preset stiffness test angle; Make the antagonistic muscle groups corresponding to the target joints co-contract according to the target tension level; Applying a corresponding thrust pulse to the target joint according to a preset disturbance torque through the flexible pneumatic joint unit, collecting the joint angle change of the target joint, and collecting the electromyographic signals corresponding to the antagonistic muscle group; Based on the disturbance torque and the joint angle change, a corresponding measurement stiffness is obtained.

8. A control device for a flexible pneumatic joint unit, the flexible pneumatic joint unit comprising at least one antagonistic pneumatic muscle pair; It is characterized in that The joint on which the flexible pneumatic joint unit is worn is used as a target joint, the antagonistic pneumatic muscle pair corresponds to the antagonistic muscle group at the target joint, and the flexible pneumatic joint unit is controlled based on the electromyographic signal of the antagonistic muscle group at the target joint, including: an intention decoupling module, configured to calculate the difference of the myoelectric signals of the antagonistic muscle groups to obtain a corresponding torque intention; further configured to calculate the sum of the myoelectric signals of the antagonistic muscle groups to obtain a corresponding stiffness intention; further configured to generate a corresponding expected torque based on the torque intention; further configured to generate a corresponding expected stiffness based on the stiffness intention; a calculation module, configured to generate a target pressure differential corresponding to each antagonistic pneumatic muscle pair based on the desired torque, and further configured to generate a base pressure corresponding to each antagonistic pneumatic muscle pair based on the desired stiffness; and further configured to generate a target pressure for each pneumatic muscle in the antagonistic pneumatic muscle pair based on the target pressure differential and the base pressure; The control module is configured to adjust the internal air pressure of the corresponding pneumatic muscle based on the target pressure.

9. The control device for the flexible pneumatic joint unit according to claim 8, characterized in that: The intention decoupling module obtains the corresponding gravity torque based on the joint angle, weight parameter and distance parameter corresponding to the target joint, wherein the weight parameter is the weight of the limb segment corresponding to the user's target joint, and the distance parameter is the distance from the center of mass of the limb segment to the target joint; The calculation module is configured to generate a corresponding target pressure difference based on the gravity torque and the expected torque.

10. A flexible exoskeleton system, characterized in that: It includes a control device, a driving device, a flexible pneumatic joint unit corresponding to each joint to be assisted, and a detection unit; The corresponding flexible pneumatic joint unit and the joint to be assisted corresponding to the detection unit are recorded as target joints; The flexible pneumatic joint unit includes at least one antagonistic pneumatic muscle pair; The detection unit comprises: The myoelectric detection component is used to collect the myoelectric signals corresponding to the antagonistic muscle group corresponding to the target joint; An air pressure detection component for collecting the internal air pressure corresponding to each pneumatic muscle in each antagonistic pneumatic muscle pair; The control device is connected to each electromyographic detection component and each air pressure detection component, and is used to execute a control method for a flexible pneumatic joint unit as described in any one of claims 1 to 7, and adjust each first pneumatic muscle and the internal air pressure of each first pneumatic muscle through a driving device.

Images