Measurement method of lower limb exoskeleton interaction force and related equipment

Abstract

The invention provides a lower limb exoskeleton interaction force measuring method and related equipment, and is applied to the technical field of data processing. The method comprises the following steps: acquiring a linear spring deformation length, a deflection angle, a spring deflection, an external joint torque signal and a target motor torque; the deformation length and the deflection angle of the linear spring are processed based on a preset neural network model, and lower limb exoskeleton static interaction force and estimated motor torque are generated; processing the external joint torque signal based on a preset neural network model to generate excitation torque; generating a lower limb exoskeleton passive power value; processing the excitation torque, the lower limb exoskeleton passive power value and the lower limb exoskeleton static interaction force based on a preset neural network model to generate an exoskeleton measurement value; processing the target motor torque and the estimated motor torque to generate a motor torque estimated error value; and processing a preset neural network model based on the motor torque pre-estimated error value to generate a lower limb exoskeleton dynamic interaction force.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method for measuring the interaction force of a lower limb exoskeleton and related equipment. Background Art

[0002] As a wearable robot, exoskeleton can enhance human functions and assist the wearer to complete tasks while providing functions such as protection and support. Lower limb exoskeleton can provide joint assistance for human walking. In terms of wearing method, flexible straps are mainly used to connect the exoskeleton and the human body, and the auxiliary force or torque is applied to the human body through a rigid binding structure. Poor interaction force will lead to low assistance efficiency and cause harm to the wearer. Therefore, the interaction force detection of exoskeleton is the key to the design of exoskeleton bionic mechanism, control strategy research, and ergonomics evaluation. Because the exoskeleton has multiple contact points with the user, it becomes difficult to isolate and measure the force transmitted between the human and the exoskeleton.

[0003] In the prior art, patent document 115847490A discloses that the user wears an exoskeleton and applies a minimum joint torque to measure the interaction force of the exoskeleton under passive dynamic conditions. It assumes that the user does not contribute any force to the movement of the exoskeleton at all, but in fact, since standing requires leg support, this assumption is unrealistic. A single-degree-of-freedom exoskeleton can use a spring-damper system to simulate the interaction between a person and an exoskeleton, thereby obtaining a single dynamic equation describing the behavior of the human body and the exoskeleton. However, it is more complicated for an exoskeleton with multiple degrees of freedom to simultaneously obtain the dynamic equations of the human body and the exoskeleton. Since different users have personalized movement patterns, identification must be repeated for each user. Therefore, how to accurately measure the interaction force of the lower limb exoskeleton is a problem that those skilled in the art need to solve.

[0004] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0005] The purpose of the present application is to provide a method for measuring the interaction force of a lower limb exoskeleton and related equipment, which at least overcomes the problems of the prior art to a certain extent, and measures and estimates the interaction force of the exoskeleton as an integral multi-body system without imposing any constraints on the exoskeleton dynamics. The interaction between the exoskeleton and the environment is limited by a precise linear spring, and the external force applied to the exoskeleton is measured by the spring length. The exoskeleton joints are stimulated by joint torque signals to imitate the human walking posture. Inverse dynamics analysis is performed to calculate the interaction torque by subtracting the exoskeleton passive power from the joint actuation torque. The measured values ​​of the exoskeleton joint inertial sensors (including joint angles and their derivatives, the angle of the thigh relative to the gravity vector and its linear acceleration) are input into an artificial neural network to estimate the interaction force of the exoskeleton. Let the user wear the exoskeleton, and use the sensor panel module and motion capture device to measure the interaction torque and gait phase between the human and the exoskeleton. The interaction force under passive dynamics estimated by the trained neural network and the contact force between the human and the exoskeleton are integrated into the feedforward controller, and the exoskeleton reference trajectory based on the interaction force is adjusted according to the feedback of the interaction torque between the human and the exoskeleton, trying to minimize the tracking error and finally converge to the expected joint trajectory.

[0006] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by the practice of the present invention.

[0007] According to one aspect of the present application, a method for measuring the interaction force of a lower limb exoskeleton is provided, comprising: obtaining a linear spring deformation length, a deflection angle, a spring deflection, an external joint torque signal, and a target motor torque; processing the linear spring deformation length and the deflection angle based on a preset neural network model to generate a lower limb exoskeleton static interaction force and an estimated motor torque; processing the external joint torque signal based on a preset neural network model to generate an excitation torque; processing the exoskeleton trunk segment position, the exoskeleton trunk segment direction, the exoskeleton right hip joint, the exoskeleton knee joint angle, the exoskeleton left hip joint, and the exoskeleton knee joint angle based on a preset neural network model to generate a lower limb exoskeleton passive power value; processing the excitation torque, the lower limb exoskeleton passive power value, and the lower limb exoskeleton static interaction force based on a preset neural network model to generate an exoskeleton measurement value; processing the target motor torque and the estimated motor torque to generate a motor torque estimation error value; processing the preset neural network model based on the motor torque estimation error value to generate a lower limb exoskeleton dynamic interaction force.

[0008] Another aspect of the present application is a device for measuring the interaction force of a lower limb exoskeleton, characterized in that the device includes: an acquisition module for acquiring a linear spring deformation length, a deflection angle, a spring deflection, an external joint torque signal, and a target motor torque; a processing module for processing the linear spring deformation length and the deflection angle based on a preset neural network model to generate a lower limb exoskeleton static interaction force and an estimated motor torque; processing the external joint torque signal based on a preset neural network model to generate an excitation torque; processing the exoskeleton trunk segment position, the exoskeleton trunk segment direction, the exoskeleton right hip joint, the exoskeleton knee joint angle, the exoskeleton left hip joint, and the exoskeleton knee joint angle based on a preset neural network model to generate a lower limb exoskeleton passive power value; processing the excitation torque, the lower limb exoskeleton passive power value, and the lower limb exoskeleton static interaction force based on a preset neural network model to generate an exoskeleton measurement value; processing the target motor torque and the estimated motor torque to generate a motor torque estimation error value; processing the preset neural network model based on the motor torque estimation error value to generate a lower limb exoskeleton dynamic interaction force.

[0009] According to another aspect of the present application, an electronic device is provided, characterized in that it comprises: a first processor; and a memory for storing executable instructions of the first processor; wherein the first processor is configured to execute the above-mentioned method for measuring the interaction force of the lower limb exoskeleton by executing the executable instructions.

[0010] According to another aspect of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a second processor, the above-mentioned method for measuring the interaction force of the lower limb exoskeleton is implemented.

[0011] According to another aspect of the present application, a computer program product is provided, comprising a computer program, wherein the computer program implements the above-mentioned method for measuring the interaction force of the lower limb exoskeleton when executed by a third processor.

[0012] The present application provides a method for measuring the interaction force of a lower limb exoskeleton and related equipment, wherein a server obtains a linear spring deformation length, a deflection angle, a spring deflection, an external joint torque signal, and a target motor torque; based on a preset neural network model, the linear spring deformation length and the deflection angle are processed to generate a lower limb exoskeleton static interaction force and an estimated motor torque; based on a preset neural network model, the external joint torque signal is processed to generate an excitation torque; based on a preset neural network model, the exoskeleton trunk segment position, the exoskeleton trunk segment direction, the exoskeleton's right hip joint, the exoskeleton's knee joint angle, the exoskeleton's left hip joint, and the exoskeleton's knee joint angle are processed to generate a lower limb exoskeleton passive power value; based on a preset neural network model, the excitation torque, the lower limb exoskeleton passive power value, and the lower limb exoskeleton static interaction force are processed to generate an exoskeleton measurement value; the target motor torque and the estimated motor torque are processed to generate a motor torque estimation error value; based on the motor torque estimation error value, the preset neural network model is processed to generate a lower limb exoskeleton dynamic interaction force. The exoskeleton dynamics are considered as a comprehensive multi-body model, taking into account the complex interactions between the exoskeleton and various parts of the user's body (such as the torso). By incorporating these interactions, the model can provide a more accurate estimate of the interaction forces and torques exchanged between the exoskeleton and the user throughout the gait cycle, and the exoskeleton passive dynamics and human-exoskeleton contact forces can improve the comfort, stability and efficiency of the exoskeleton during assisted walking.

[0013] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 A flow chart showing a method for measuring interaction force of a lower limb exoskeleton provided in one embodiment of the present application;

[0015] Figure 2 A schematic structural diagram of a device for measuring interaction forces of a lower limb exoskeleton provided in one embodiment of the present application is shown;

[0016] Figure 3 A schematic diagram of the structure of an electronic device provided by an embodiment of the present application is shown;

[0017] Figure 4 A schematic diagram of a storage medium provided by an embodiment of the present application is shown;

[0018] Figure 5 A schematic diagram of exoskeleton coordinates and joint angles of a measuring device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0019] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0020] Combine the following Figure 1 To describe the method for measuring the interaction force of the lower limb exoskeleton according to an exemplary embodiment of the present application. It should be noted that the following application scenarios are only shown to facilitate understanding of the spirit and principles of the present application, and the embodiments of the present application are not limited in this regard. On the contrary, the embodiments of the present application can be applied to any applicable scenario.

[0021] In one embodiment, the present application also proposes a method for measuring the interaction force of a lower limb exoskeleton and related equipment. Figure 1 The following schematically shows a flow chart of a method for measuring the interaction force of a lower limb exoskeleton according to an embodiment of the present application. Figure 1 As shown, the method is applied to a server, comprising:

[0022] S101, obtaining linear spring deformation length, deflection angle, spring deflection, external joint torque signal, and target motor torque.

[0023] In one embodiment, the exoskeleton is suspended from the ceiling using linear springs with known spring constant K, and sensors are installed at both ends for motion capture. The exoskeleton is suspended during the dynamic recognition data collection process. During the experiment, the joints are stimulated, and the interaction of the exoskeleton with the environment is limited to the force applied by the linear springs. The linear spring deformation length l and the deflection angle α, i.e., the spring deflection K, are captured using a high-precision camera.

[0024] In another embodiment, reference Figure 5 , Figure 5 The diagram is a schematic diagram of the exoskeleton coordinates and joint angles of the measurement device. The global, back support (trunk) and thigh coordinate systems are represented by 0, 1 and 2 respectively. The angles of the trunk and thigh relative to the gravity vector are represented by θ and γ respectively. h and q k Represent the angles of the hip joint and knee joint respectively. p1 and p2 represent the coordinate positions of the torso and thigh in the global coordinate system (x0, y0). The acceleration vector a is measured by the inertial sensor embedded in the exoskeleton thigh. The interaction force vector f between the environment and the exoskeleton int Calculated from the linear spring deflection measured from motion capture. The distances from the torso frame origin to the hip rotation axis and from the hip rotation axis to the accelerometer location on the exoskeleton thigh are represented by l1 and l2, and the gravity axis is represented by a dotted line.

[0025] S102, processing the linear spring deformation length and the deflection angle based on a preset neural network model to generate a lower limb exoskeleton static interaction force and an estimated motor torque.

[0026] In one embodiment, joint torque signals are used to stimulate the exoskeleton joints to simulate human walking posture. The selection of joint torque signals needs to cover the speed range and joint motion range of normal human walking.

[0027] In addition, the preset neural network model includes a calculation formula for obtaining the static interaction force of the lower limb exoskeleton, which is: f int =Kl; where K is the spring deflection and l is the linear spring deformation length.

[0028] In another embodiment, the preset neural network model includes a calculation formula for obtaining the estimated motor torque, which is: int =[f int ; τ int ]; τ int =f int l; where f int is the static interaction force of the lower limb exoskeleton, τ int is the interaction torque of the exoskeleton joint (force × lever arm), and l is the linear spring deformation length.

[0029] Suppose we are studying a lower limb exoskeleton device with the goal of accurately measuring its various mechanical parameters in the process of assisting human walking, especially the interaction force of the lower limb exoskeleton. The following is a specific example based on the above content to help better understand the whole process.

[0030] We use a linear spring with a known spring stiffness coefficient of K = 50N / m to suspend the exoskeleton from the ceiling. High-precision sensors are installed at both ends of the spring to capture the deformation information of the spring. The exoskeleton remains suspended during the dynamic recognition data acquisition process. At this time, the joints are subjected to specific excitations, and the interaction between the exoskeleton and the environment is only reflected by the force applied by the linear spring. Using a high-precision camera, we measured the linear spring deformation length l = 0.2m and the deflection angle (assuming α = 30°. The deflection angle here is mainly used for other possible subsequent calculations or analyses. It is not directly used in the current calculation of the static interaction force of the lower limb exoskeleton, but it is also an important monitoring parameter for the exoskeleton status).

[0031] Calculate the static interaction force of the lower limb exoskeleton (based on f int =K1). According to formula f int =Kl, substituting K = 50N / m and l = 0.2m, we can get: f int=50×0.2=10N. This means that due to the deformation of the linear spring, the exoskeleton is subjected to a static interaction force of 10N. This force is the result of the interaction between the exoskeleton and the external environment (through the spring). It reflects the tension or pressure from the spring that the exoskeleton is subjected to in the current suspension state, and is an important basic data for the subsequent analysis of the mechanical state of the exoskeleton.

[0032] Calculate the estimated motor torque (based on u int =[f int ; τ int ]). First, calculate the interaction torque τ of the exoskeleton joints int , assuming that the exoskeleton is designed so that the arm length (related to the linear spring deformation length, here assumed to be l = 0.2m, the actual situation may be determined according to the specific structure of the exoskeleton) is 0.2m, according to τ int =f int l, we can get: τ int =10×0.2=2N.m. Then calculate the estimated motor torque u int , f int =10N and τ int =2N.m Substitute into the formula u int =[f int ; τ int ], get u int =[10N; 2N.m].

[0033] Here the motor torque u is estimated int It contains information about the static interaction force of the lower limb exoskeleton and the interaction torque of the exoskeleton joints. It is the motor torque estimated based on the static state of the current exoskeleton (spring deformation, etc.), which provides an important basis for further analysis of the dynamic characteristics of the exoskeleton. For example, when designing the motor control system of the exoskeleton, this estimated motor torque can be used as a reference value to help determine the output torque range of the motor to achieve precise control of the exoskeleton movement.

[0034] In the actual exoskeleton-assisted walking process, accurate measurement and estimation of these parameters are crucial to optimizing the performance of the exoskeleton. For example, when a person walks in an exoskeleton, the exoskeleton needs to adjust the motor output torque in real time according to factors such as the body's motion state and ground reaction force to provide appropriate assistance. The lower limb exoskeleton static interaction force and estimated motor torque calculated by this method can help the exoskeleton system better understand the mechanical relationship between itself, the environment, and the human body, thereby providing more accurate and comfortable assistance in different walking stages (such as starting, acceleration, constant speed, deceleration, etc.), improving walking stability and efficiency, and reducing adverse effects on the human body (such as discomfort caused by excessive assistance or insufficient assistance affecting walking effects, etc.). Through this example, we can clearly see the importance and practical application value of the entire calculation process and these parameters in exoskeleton mechanics analysis and control.

[0035] S103, processing the external joint torque signal based on a preset neural network model to generate an excitation torque.

[0036] In one embodiment, the preset neural network model includes a calculation formula for obtaining the excitation torque, which is: act =[0 3×1 ; τ act ]; where u act is the external joint torque signal, including the motor torque of the hip joint and knee joint of both legs, 0 3×1 represents a 3*1 zero matrix, τ act Represents the excitation torque.

[0037] Assuming that we are studying a lower limb exoskeleton system, the following is an example of obtaining the formula for calculating the excitation torque to help understand its application in practical scenarios. Consider a lower limb exoskeleton with power assist functions for the hip and knee joints of both legs. In the experiment, we want to simulate the normal walking posture of the human body and analyze the mechanical characteristics of the exoskeleton under different motion states, especially the effect of the excitation torque on its motion. The control system of the exoskeleton can acquire and process various sensor data to achieve precise control of the exoskeleton joints. Among them, the external joint torque signal is a key input parameter, which reflects the torque applied to the joint by the outside, and is of great significance for studying the dynamic response and motion control strategy of the exoskeleton.

[0038] Calculate the excitation torque (based on u act =[0 3×1 ; τ act]). Assume that at a certain moment, the motor torques of the hip and knee joints of both legs are measured by the exoskeleton sensors as follows: the motor torque of the right hip joint is 5 N.m, the motor torque of the right knee joint is 3 N.m, the motor torque of the left hip joint is 4 N.m, and the motor torque of the left knee joint is 2 N.m. Then, the external joint torque signal u act τ in act The part can be expressed as [5N.m; 3N.m; 4N.m; 2N.m] T (This represents transposition, and the column vector form is expressed so that it corresponds to the form in the formula.) 3×1 is a zero matrix. In this example, its role in the formula is to distinguish different parts of information in the vector representation or to meet specific computational structure requirements. Therefore, according to the formula u act =[0 3×1 ; τ act ], the excitation torque at this time is [0; 0; 0; 5N.m; 3N.m; 4N.m; 2N.m] T .

[0039] The calculated excitation torque u act It will be input into the control system of the exoskeleton or further calculation model. For example, when analyzing the dynamic characteristics of the exoskeleton, the excitation torque will be calculated together with other parameters (such as the inertia and friction of the exoskeleton) to determine the changes in the motion state of the exoskeleton joints, such as acceleration, velocity and displacement. In the actual walking process, different walking stages (such as starting, accelerating, turning, etc.) require excitation torques of different sizes and directions to drive the movement of the exoskeleton joints to assist the human body in completing the corresponding actions. By accurately acquiring and processing the excitation torque, the exoskeleton can better adapt to the needs of human movement, provide appropriate assistance, and improve the comfort and efficiency of walking. For example, in the starting stage, a large excitation torque may be required to overcome static friction and start the movement of the exoskeleton joints; while when walking at a uniform speed, the excitation torque is mainly used to maintain the stable motion state of the joints and compensate for possible energy losses.

[0040] Excitation torque u act It is closely related to other parameters of the exoskeleton. For example, it is related to the calculation of the passive dynamic value (I′) of the lower limb exoskeleton. When calculating I′ later, u act As an important input parameter, it is related to the estimated motor torque (u int) and other factors jointly affect the value of Γ, which in turn affects the overall dynamic analysis results of the exoskeleton. At the same time, accurate calculation of the excitation torque is crucial for optimizing the exoskeleton's control strategy. By continuously adjusting the size and direction of the excitation torque, combined with real-time monitoring of the exoskeleton's motion state and human feedback information, a more intelligent and precise control of the exoskeleton can be achieved, enabling it to better assist the human body in completing various tasks while reducing energy consumption and adverse effects on the human body (such as discomfort caused by excessive or insufficient assistance). Through this example, we can more clearly understand the specific application and significance of the excitation torque calculation formula in the lower limb exoskeleton system, and how it interacts with other parameters to affect the overall performance of the exoskeleton.

[0041] S104, based on a preset neural network model, the exoskeleton trunk segment position, the exoskeleton trunk segment direction, the exoskeleton right hip joint, the exoskeleton knee joint angle, the exoskeleton left hip joint, and the exoskeleton knee joint angle are processed to generate a lower limb exoskeleton passive power value.

[0042] In one embodiment, the preset neural network model includes a calculation formula for obtaining the passive power value of the lower limb exoskeleton, and the calculation formula is:

[0043] I′(θ, θ', θ″, p1, q, q′, q″)=u act +u int ; p1=(p 1，x ;p 1，y );

[0044] q=(q h，r ;q k,r ;q h,l ;q k，l ), where I′ represents the passive dynamic value of the lower limb exoskeleton, θ represents the direction of the trunk segment of the exoskeleton, θ′ and θ″ represent the first-order derivative and second-order derivative of θ, respectively, and p1 = (p 1，x ;p 1,y ) represents the position of the exoskeleton’s trunk segment, q=(q h，r ;q k,r ;q h,l ;q k，l ) represents the right hip joint and knee joint angle, left hip joint and knee joint angle of the exoskeleton, q′ and q″ represent the first-order derivative and second-order derivative of q, respectively, u act Represents the external joint torque signal, u int Represents the estimated motor torque.

[0045] The following is an example of the formula for calculating the passive dynamic value of a lower limb exoskeleton to help understand its application in actual situations. Suppose a user is wearing a lower limb exoskeleton for a walking experiment, and we want to analyze the passive dynamic value of the exoskeleton at this time. Through the inertial sensors and angle sensors on the exoskeleton and other devices, the following data is obtained: the direction of the exoskeleton's trunk segment θ = 45° (for the convenience of the example, assume that a certain reference direction is 0°, and the angle here represents the orientation of the trunk segment relative to the reference direction), its first-order derivative θ' = 0.5rad / s (indicating the rate of change of the trunk segment direction), and the second-order derivative θ'′ = 0.1rad / s 2 (Indicates the change in the rate of change of direction). The exoskeleton trunk segment position p1 = (1.0m; 0.5m) (Here it is assumed that it is in a two-dimensional plane, p 1，x =1.0m represents the position coordinate in the x direction, p 1，y =0.5m represents the position coordinate in the y direction).

[0046] Exoskeleton joint angles: right hip joint angle q h，r =30°, right knee joint angle q k，r =60°, left hip joint angle q h,l =25. , left knee joint angle q k,l =55°, and the corresponding first-order derivatives are q′ h，r =0.3rad / s,q′ k，r =0.4rad / s,q′ h，l =0.25rad / s,q′ k，l =0.35rad / s (indicates the rate of change of joint angle), and the second-order derivatives are q″ h，r =0.05rad / s 2 , q″ k，r =0.08rad / s 2 , q″ h，l =0.04rad / s 2 , q″ k，l =0.065rad / s 2 (Indicates the change in the rate of change of the joint angle). External joint torque signal u act (Assuming that it has been obtained through measurement and calculation, such as [0; 0; 0; 5N.m; 3N.m; 4N.m; 2N.m] obtained in the previous excitation torque calculation example T , here simplified to [5N.m; 3N.m; 4N.m; 2N.m] T denoted), estimated motor torque u int [8N.m; 3N.m] T (Also assumed to have been obtained through previous calculations or other means).

[0047] Calculate the passive power value (I′) of the lower limb exoskeleton according to the formula I′(θ, θ′, θ″, p1, q, q′, q″) = u act +u int , substitute the above data into the formula. First, add the corresponding data in vector form:

[0048] u act +u int =[5N.m; 3N.m; 4N.m; 2N.m] T +[8N.m; 3N.m] T =[13N; 6N.m; 4N.m; 2N.m] T .

[0049] The calculated passive power value of the lower limb exoskeleton is I′=[13N; 6N.m; 4N.m; 2N.m] T reflects the passive power of the exoskeleton in its current state. Each element in the vector may correspond to a different physical meaning. For example, the first element 13N may be related to the overall linear force, while the following 6N.m, 4N.m, 2N.m, etc. are related to the torque at different joints (depending on the mechanical model definition of the exoskeleton). This passive power value is very important for understanding the performance of the exoskeleton during assisted walking. For example, when designing the control strategy of the exoskeleton, the output of the motor can be adjusted according to the value of to better assist human movement. If a value in is too large or too small, it may mean that the mechanical state of the exoskeleton in a certain aspect is not ideal, and corresponding adjustments need to be made, such as optimizing the joint structure, adjusting the motor torque distribution, etc., to improve the comfort, stability and efficiency of the exoskeleton. At the same time, through real-time monitoring of changes, problems that may occur during the operation of the exoskeleton, such as abnormal increase in resistance or insufficient assistance, can be discovered in a timely manner.

[0050] S105, processing the excitation torque, the passive dynamic value of the lower limb exoskeleton and the static interaction force of the lower limb exoskeleton based on a preset neural network model to generate an exoskeleton measurement value.

[0051] In one embodiment, the preset neural network model includes a calculation formula for obtaining the exoskeleton measurement value, and the calculation formula is:

[0052] x=I ′-1 (τ act , f int );

[0053] Among them, I′ is the passive power value of the lower limb exoskeleton, τ act is the excitation torque, f int is the static interaction force of the lower limb exoskeleton.

[0054] The following is an example of an exoskeleton measurement calculation formula to help understand its practical application in a lower limb exoskeleton system. Assume that a lower limb exoskeleton is being tested for performance, and during the test, we have calculated or measured some key parameters through the previous steps.

[0055] It is known that the passive dynamic value of the lower limb exoskeleton is I′=[10N; 5N.m; 3N.m; 1N.m] T (Here, the vector form is assumed, and each element corresponds to different physical meanings, such as linear force and torque at different joints, etc.), the excitation torque T act =[2N.m; 1N.m; 0.5Nm; 0.2Nm] T , static interaction force of lower limb exoskeleton f int =8N.

[0056] Calculate exoskeleton measurements (x):

[0057] According to the formula x = I' -1 (τ act , f int ), where I′ -1 It represents the inverse operation of I′, and its specific form may vary depending on the specific mechanical model and calculation method of the exoskeleton system.

[0058] Assume that in this particular exoskeleton model, I′ -1 The calculation method is: for vector I' = [I'1; I'2; I'3; I'4] T , I′ -1 (τ act , f int ) is calculated as a scalar x, and the calculation formula is

[0059] Substitute the known data into the formula to calculate:

[0060]

[0061] The calculated exoskeleton measurement value x = 3.3375 is of great significance in the analysis of the entire exoskeleton system. It integrates multiple key factors such as the passive power value, excitation torque and static interaction force of the lower limb exoskeleton, and may be used to evaluate the current working state, performance or degree of deviation from the expected target of the exoskeleton. For example, when evaluating the power-assisting effect of the exoskeleton, if the value deviates from the preset ideal range, it may mean that the power-assisting strategy of the exoskeleton needs to be adjusted, such as changing the size of the excitation torque or adjusting the distribution of passive power. At the same time, by monitoring and analyzing the x value at multiple different times or under different motion states, the dynamic performance changes of the exoskeleton during the entire working process can be understood, providing a basis for further optimizing the exoskeleton design and control algorithm. In addition, the value may also be related to the safety assessment of the exoskeleton. If the value is abnormal, it may indicate that there are potential risks in the exoskeleton system, such as excessive force or insufficient power, which need to be checked and adjusted in time.

[0062] S106: Process the target motor torque and the estimated motor torque to generate a motor torque estimation error value.

[0063] In one embodiment, assume that there is a lower limb exoskeleton test experiment, and the subject walks in a specific venue after wearing the exoskeleton. We assume that a set of exoskeleton measurement values ​​x have been obtained through the previous steps, where the joint angle q = (30°, 60°, 25°, 55°) (corresponding to the right hip joint, right knee joint, left hip joint, and left knee joint angles, respectively), its first-order derivative q′ = (0.3rad / s, 0.4rad / s, 0.25rad / s, 0.35rad / s), and the second-order derivative q″ = (0.05rad / s 2 , 0.08rad / s 2 , 0.04rad / s 2 , 0.06rad / s 2 ); The angle of the thigh relative to the gravity vector is γ = 45°, and its linear acceleration p1 = (1.5 m / s 2 , 0.8m / s 2 )(assuming two directional components in a two-dimensional plane).

[0064] These data are input into the neural network model, and the estimated interaction force f′ is obtained through model calculation. int = 12N and estimated motor torque T′ act =[4N.m; 2N.m; 3N.m; 1N.m] T At the same time, the target motor torque (ie, the actual motor torque) T is obtained through the actual measurement device. act =[3.5Nm; 1.8Nm; 2.5Nm; 0.9Nm] T .

[0065] The motor torque estimation error can be calculated in many ways. Here we assume a simple difference calculation method, that is, each corresponding element is subtracted to obtain the error vector.

[0066] The error vector is calculated as follows: ΔT act =T′ act -T act =[4N.m-3.5Nm; 2N.m-1.8Nm; 3N.m-2.5Nm; 1N.m-0.9Nm].

[0067] This error vector reflects the deviation between the estimated motor torque and the actual motor torque at each joint. For example, for the right hip joint, the estimated torque is 0.5Nm greater than the actual torque, which may mean that the neural network model has a certain deviation in the torque estimation of this joint, and it is necessary to further adjust the model parameters or improve the measurement method to improve the accuracy of the estimation.

[0068] To obtain actual exoskeleton measurements, the sensor panel module is installed on the rigid extension structure, such as the key parts of the exoskeleton's legs and waist. When the subject walks, the flexible capacitive sensor body collects the capacitance data, and the acquisition circuit packages it and sends it to the data processing and analysis system.

[0069] Assume that at a certain moment, the contact torque data obtained after the conversion of the collected capacitance data is T contact =[0.8Nm; 0.3Nm; 0.6Nm; 0.2Nm] T (corresponding to the contact torque of different parts respectively), and the motion capture device detects the spatial position of the exoskeleton (1.2m, 0.6m, 0.5m) (assuming three-dimensional spatial coordinates), speed (0.5m / s, 0.3m / s, 0.2m / s), acceleration (0.8m / s 2 , 0.4m / s 2 , 0.3m / s 2 ), angles (35°, 65°, 28°, 58°), angular velocities (0.35rad / s, 0.45rad / s, 0.28rad / s, 0.38rad / s) and angular acceleration (0.06rad / s 2 , 0.09rad / s 2 , 0.05rad / s 2 , 0.07rad / s 2 ), and determine the specific phase of the gait cycle (e.g., position 20% of the starting phase) based on gait events such as heel strike.

[0070] These data together constitute the actual exoskeleton measurement value X, which contains rich status information of the exoskeleton during actual use, including not only the contact torque related to the motor torque, but also other parameters reflecting the overall motion state of the exoskeleton and the human-computer interaction.

[0071] The relevant data (such as contact torque, joint angle, etc.) in the actual exoskeleton measurement value X are input into the neural network model again to re-estimate the motor torque T″ act (Assuming that T″ act =[3.6Nm; 1.9Nm; 2.6Nm; 1.1Nm] T ).

[0072] Then the motor torque estimation error is calculated.

[0073] ΔT′ act =T″ act -T act =[3.6Nm-3.5Nm; 1.9Nm-1.8Nm; 2.6Nm-2.5Nm; 1.1Nm-0.9Nm] T .

[0074] By comparing the motor torque estimation error values ​​obtained from the two implementations, we can find out the impact of different measurement and calculation methods on the results. For example, in this example, the error value calculated based on the more comprehensive actual exoskeleton measurement value X is relatively small, indicating that using the sensor panel module and the motion capture device to obtain more actual data helps to improve the accuracy of motor torque estimation, which can provide a more reliable basis for the control and optimization of the exoskeleton. At the same time, based on these error values, the performance of the exoskeleton under different working conditions can be further analyzed, such as whether there is a large torque estimation error in the starting stage, so as to improve the design and control algorithm of the exoskeleton in a targeted manner.

[0075] S107, processing the preset neural network model based on the motor torque estimation error value to generate a lower limb exoskeleton dynamic interaction force.

[0076] In one implementation, assume that we are testing and optimizing the performance of a lower limb exoskeleton, and the motor torque estimation error value has been obtained in the previous calculation. For example, the motor torque estimation error value ΔT calculated previously act =[0.5Nm; 0.2Nm; 0.5Nm; 0.1Nm] T (corresponding to the hip and knee joints of both legs), and the previously estimated motor torque T′ is known act =[4N.m; 2N.m; 3N.m; 1N.m] T , the actual motor torque T act=[3.5Nm; 1.8Nm; 2.5Nm; 0.9Nm] T .

[0077] By replacing the estimated motor torque with the actual motor torque, we obtain new data for calculating the dynamic interaction force. act Participate in subsequent calculations (assuming that the formula for calculating the dynamic interaction force is F dynamic = k(T act , f int ), where k is a function related to the structure and mechanical properties of the exoskeleton).

[0078] Assume that in the current exoskeleton state, the lower limb exoskeleton static interaction force f int = 10N, and according to the specific mechanical model and calculation method of the exoskeleton (assuming that the calculation method of the function is The dynamic interaction force F is calculated dynamic =10N×(3.5N.m+1.8N.m+2.5N.m+0.9Nm)=87N.

[0079] Next, the estimated error is used to update the neural network weights. Assume that the current neural network weight vector is W = [0.2, 0.3, 0.1, 0.4] (this is a simplified example, the actual weight vector dimension may be higher and related to more input parameters), and the learning rate α = 0.05 (the learning rate is a hyperparameter that can be adjusted based on experience or experiments).

[0080] According to the formula, the new weight W new =W-α×ΔT act , calculate the updated weight vector:

[0081] W new =[0.2, 0.3, 0.1, 0.4]-0.05×[0.5Nm, 0.2Nm, 0.5Nm, 0.1Nm] T =[0.175, 0.29, 0.075, 0.395] T .

[0082] The lower limb exoskeleton dynamic interaction force F is obtained by replacing the estimated motor torque with the actual motor torque dynamic= 87N, which reflects the dynamic interaction between the exoskeleton and the external environment (including the human body) under the actual motor torque. In practical applications, this value can be used to evaluate the mechanical properties of the exoskeleton in the current motion state, such as judging whether the assistance is appropriate, whether there is abnormal interaction force, etc. If the dynamic interaction force is too large or too small, the control strategy or structural design of the exoskeleton may need to be further adjusted.

[0083] Updated neural network weights W new =[0.175, 0.29, 0.075, 0.395] T , which means that the neural network has adjusted its own parameters based on the estimation error this time. As more data is input and the weights are updated, the estimation accuracy of the neural network is expected to improve. For example, in subsequent calculations, using the updated weights, for the same or similar input data (such as exoskeleton measurements), the estimated motor torque and interaction force will be closer to the actual values, so that the exoskeleton can better adapt to different motion scenarios, provide more accurate assistance, and improve the user experience and the overall performance of the exoskeleton. This example shows how to improve the calculation of the exoskeleton dynamic interaction force and optimize the neural network model based on the motor torque estimation error value, thereby achieving the improvement of exoskeleton performance and optimization of control. Different exoskeleton systems may have different specific calculation methods and parameter settings, but the overall idea and process are similar.

[0084] In this application, the server obtains the linear spring deformation length, deflection angle, spring deflection, external joint torque signal, and target motor torque; based on the preset neural network model, the linear spring deformation length and deflection angle are processed to generate the static interaction force of the lower limb exoskeleton and the estimated motor torque; based on the preset neural network model, the external joint torque signal is processed to generate the excitation torque; based on the preset neural network model, the exoskeleton trunk segment position, the exoskeleton trunk segment direction, the exoskeleton right hip joint, the exoskeleton knee joint angle, the exoskeleton left hip joint, and the exoskeleton knee joint angle are processed to generate the lower limb exoskeleton passive power value; based on the preset neural network model, the excitation torque, the lower limb exoskeleton passive power value and the lower limb exoskeleton static interaction force are processed to generate the exoskeleton measurement value; the target motor torque and the estimated motor torque are processed to generate the motor torque estimation error value; based on the motor torque estimation error value, the preset neural network model is processed to generate the lower limb exoskeleton dynamic interaction force. The exoskeleton dynamics is regarded as a comprehensive multi-body model, taking into account the complex interaction between the exoskeleton and various parts of the user's body (such as the torso). By incorporating these interactions, the model can provide a more accurate estimate of the interaction forces and torques exchanged between the exoskeleton and the user throughout the gait cycle, combining exoskeleton passive dynamics and human-exoskeleton contact forces to improve the comfort, stability, and efficiency of the exoskeleton during assisted walking.

[0085] Optionally, in another embodiment of the above method of the present application, after the preset neural network model is processed based on the motor torque estimation error value to generate the lower limb exoskeleton dynamic interaction force, the method further includes:

[0086] Obtain the contact force between the user and the exoskeleton;

[0087] A preset joint assistance trajectory is generated based on the contact force between the user and the exoskeleton and the dynamic interaction force of the lower limb exoskeleton.

[0088] In one embodiment, the measured values ​​(joint angles and their derivatives, thigh angles relative to the gravity vector and their linear acceleration) and motor torque measured by the motion capture device are input into the trained neural network model to obtain the interaction force under the passive dynamic state of the exoskeleton. The interaction force under the passive dynamic state and the human-exoskeleton contact force are integrated into the feedforward controller, and the exoskeleton reference trajectory based on the interaction force is adjusted according to the feedback of the human-computer interaction control, in an attempt to minimize the tracking error, and finally converge to the joint trajectory expected by humans.

[0089] Since the exoskeleton is used to assist the user, it cannot "follow" the user's trajectory too closely (which will not provide assistance), nor can it deviate too far from the user's trajectory to make the user feel uncomfortable (don't try too hard). Through feedback, the error between the exoskeleton's reference trajectory and the user's trajectory is minimized and converged to an expected assistance trajectory. The user attempts to control the position of his joints on his predetermined trajectory, while the exoskeleton, without knowing the user's predetermined trajectory, attempts to minimize the error between its position and the reference trajectory.

[0090] By applying the above technical solutions, the exoskeleton is regarded as an integral multi-body system, and the interaction force of the exoskeleton is measured and estimated without imposing any constraints on the exoskeleton dynamics. The interaction between the exoskeleton and the environment is limited by precise linear springs, and the external force applied to the exoskeleton is measured by the spring length. The exoskeleton joints are stimulated by joint torque signals to imitate the human walking posture. Inverse dynamics analysis is performed to subtract the exoskeleton passive power from the joint actuation torque to calculate the interaction torque. The measurement values ​​of the exoskeleton joint inertial sensors (including joint angles and their derivatives, the angle of the thigh relative to the gravity vector and its linear acceleration) are input into the artificial neural network to estimate the interaction force of the exoskeleton. Let the user wear the exoskeleton, and use the sensor panel module and motion capture device to measure the interaction torque and gait phase between the human and the exoskeleton. The interaction force under passive dynamics estimated by the trained neural network and the human-exoskeleton contact force are integrated into the feedforward controller. The exoskeleton reference trajectory based on the interaction force is adjusted according to the feedback of the interaction torque between the human and the exoskeleton, trying to minimize the tracking error and finally converge to the expected joint trajectory.

[0091] In one embodiment, if Figure 2 As shown, the present application also provides a device for measuring the interaction force of a lower limb exoskeleton, comprising:

[0092] An acquisition module 201 is used to acquire linear spring deformation length, deflection angle, spring deflection, external joint torque signal, and target motor torque;

[0093] The processing module 202 is used to process the linear spring deformation length and the deflection angle based on a preset neural network model to generate a lower limb exoskeleton static interaction force and an estimated motor torque; process the external joint torque signal based on a preset neural network model to generate an excitation torque; process the exoskeleton trunk segment position, the exoskeleton trunk segment direction, the exoskeleton right hip joint, the exoskeleton knee joint angle, the exoskeleton left hip joint, and the exoskeleton knee joint angle based on a preset neural network model to generate a lower limb exoskeleton passive power value; process the excitation torque, the lower limb exoskeleton passive power value and the lower limb exoskeleton static interaction force based on a preset neural network model to generate an exoskeleton measurement value; process the target motor torque and the estimated motor torque to generate a motor torque estimation error value; process the preset neural network model based on the motor torque estimation error value to generate a lower limb exoskeleton dynamic interaction force.

[0094] The present application embodiment provides an electronic device, such as Figure 3 As shown, the electronic device 3 includes a first processor 300, a memory 301, a bus 302 and a communication interface 303, wherein the first processor 300, the communication interface 303 and the memory 301 are connected via the bus 302; the memory 301 stores a computer program that can be run on the first processor 300, and when the first processor 300 runs the computer program, the method for measuring the interaction force of the lower limb exoskeleton provided in any of the aforementioned embodiments of the present application is executed.

[0095] The memory 301 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 303 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.

[0096] The bus 302 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The memory 301 is used to store programs, and the first processor 300 executes the programs after receiving the execution instruction. The method for measuring the interaction force of the lower limb exoskeleton disclosed in any implementation of the embodiment of the present application may be applied to the first processor 300, or implemented by the first processor 300.

[0097] The first processor 300 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the first processor 300. The above-mentioned first processor 300 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a readily available programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be embodied as a hardware decoding processor to be executed, or a combination of hardware and software modules in the decoding processor to be executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 301, and the first processor 300 reads the information in the memory 301 and completes the steps of the above method in combination with its hardware.

[0098] The electronic device provided in the above-mentioned embodiments of the present application and the method for measuring the interaction force of the lower limb exoskeleton provided in the embodiments of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.

[0099] The present application embodiment provides a computer-readable storage medium, such as Figure 4 As shown, the computer-readable storage medium 401 stores a computer program, and when the computer program is read and executed by the second processor 402, the aforementioned method for measuring the interaction force of the lower limb exoskeleton is implemented.

[0100] The technical solution of the embodiment of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling an electronic device (which may be an air conditioner, a refrigeration device, a personal computer, a server, or a network device, etc.) or a processor to execute all or part of the steps of the method described in the embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a ROM, a RAM, a magnetic disk, or an optical disk.

[0101] The computer-readable storage medium provided in the above-mentioned embodiment of the present application and the method for measuring the interaction force of the lower limb exoskeleton provided in the embodiment of the present application are based on the same inventive concept and have the same beneficial effects as the method adopted, run or implemented by the application program stored therein.

[0102] The embodiment of the present application provides a computer program product, including a computer program, which is executed by a third processor to implement the method described above. The computer program product provided by the above embodiment of the present application and the method for measuring the interaction force of the lower limb exoskeleton provided by the embodiment of the present application are based on the same inventive concept and have the same beneficial effects as the method adopted, run or implemented by the application program stored therein.

[0103] Each embodiment in the present application is described in a related manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the measurement method, electronic device, electronic device, and readable storage medium embodiment for evaluating the interaction force of a lower limb exoskeleton, since they are basically similar to the above-mentioned embodiment of the measurement method for the interaction force of a lower limb exoskeleton, the description is relatively simple, and the relevant parts can be referred to the partial description of the above-mentioned embodiment of the measurement method for the interaction force of a lower limb exoskeleton.

[0104] Although the present application is disclosed as above, the present application is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present application. Therefore, the protection scope of the present application shall be subject to the scope defined by the claims.

Claims

1. A method for measuring the interaction force of a lower limb exoskeleton, characterized in that: include: Obtain linear spring deformation length, deflection angle, spring deflection, external joint torque signal, and target motor torque; Processing the linear spring deformation length and the deflection angle based on a preset neural network model to generate a lower limb exoskeleton static interaction force and an estimated motor torque; Processing the external joint torque signal based on a preset neural network model to generate an excitation torque; Based on the preset neural network model, the exoskeleton trunk segment position, the exoskeleton trunk segment direction, the exoskeleton right hip joint, the exoskeleton knee joint angle, the exoskeleton left hip joint, and the exoskeleton knee joint angle are processed to generate the lower limb exoskeleton passive power value; Processing the excitation torque, the passive dynamic value of the lower limb exoskeleton and the static interaction force of the lower limb exoskeleton based on a preset neural network model to generate an exoskeleton measurement value; Processing the target motor torque and the estimated motor torque to generate a motor torque estimation error value; The preset neural network model is processed based on the motor torque estimation error value to generate a lower limb exoskeleton dynamic interaction force.

2. The method according to claim 1, characterized in that The method processes the linear spring deformation length and the deflection angle based on a preset neural network model to generate a lower limb exoskeleton static interaction force and an estimated motor torque, including: The preset neural network model includes a calculation formula for obtaining the static interaction force of the lower limb exoskeleton, and the calculation formula is: f int =Kl; Where K is the spring deflection and l is the linear spring deformation length.

3. The method according to claim 2, characterized in that The method further comprises: processing the linear spring deformation length and the deflection angle based on a preset neural network model to generate a lower limb exoskeleton static interaction force and an estimated motor torque; The preset neural network model includes a calculation formula for obtaining the estimated motor torque, and the calculation formula is: you int =[f int ;t int ]; t int =f int l; Among them, f int is the static interaction force of the lower limb exoskeleton, τ int is the interaction torque of the exoskeleton joint (force × lever arm), and l is the linear spring deformation length.

4. The method according to claim 3, characterized in that The processing of the external joint torque signal based on a preset neural network model to generate an excitation torque includes: The preset neural network model includes a calculation formula for obtaining the excitation torque, and the calculation formula is: you act =[0 3×1 ;t act ]; Among them, u act is the external joint torque signal, including the motor torque of the hip joint and knee joint of both legs, 0 3×1 represents a 3*1 zero matrix, τ act Represents the excitation torque.

5. The method according to claim 4, characterized in that The method processes the exoskeleton trunk segment position, the exoskeleton trunk segment direction, the exoskeleton right hip joint, the exoskeleton knee joint angle, the exoskeleton left hip joint, and the exoskeleton knee joint angle based on the preset neural network model to generate the lower limb exoskeleton passive power value, including: The preset neural network model includes a calculation formula for obtaining the passive power value of the lower limb exoskeleton, and the calculation formula is: I′(θ,θ',θ″,p1,q,q′,q″)=u act +u int ; p1=(p 1，x ;p 1，y ); q=(q h，r ;q k，r ;q h，l ;q k，l ); Where I′ represents the passive dynamic value of the lower limb exoskeleton, θ represents the direction of the exoskeleton trunk segment, θ′ and θ″ represent the first-order derivative and second-order derivative of θ, p1=(p1, x; p1, y) represents the position of the exoskeleton trunk segment, and q=(q h，r ;q k，r ;q h，l ;q k，l ) represents the right hip joint and knee joint angle, left hip joint and knee joint angle of the exoskeleton, q′ and q″ represent the first-order derivative and second-order derivative of q, respectively, u act Represents the external joint torque signal, u int Represents the estimated motor torque.

6. The method according to claim 5, characterized in that The step of processing the excitation torque, the passive dynamic value of the lower limb exoskeleton and the static interaction force of the lower limb exoskeleton based on a preset neural network model to generate an exoskeleton measurement value includes: The preset neural network model includes a calculation formula for obtaining the exoskeleton measurement value, and the calculation formula is: x=I′ -1 (t act ,f int ); Among them, I′ is the passive power value of the lower limb exoskeleton, τ act is the excitation torque, f int is the static interaction force of the lower limb exoskeleton.

7. The method according to claim 1, characterized in that After the preset neural network model is processed based on the motor torque estimation error value to generate the lower limb exoskeleton dynamic interaction force, the method further includes: Obtain the contact force between the user and the exoskeleton; A preset joint assistance trajectory is generated based on the contact force between the user and the exoskeleton and the dynamic interaction force of the lower limb exoskeleton.

8. A device for measuring the interaction force of a lower limb exoskeleton, characterized in that: The device comprises: An acquisition module is used to acquire linear spring deformation length, deflection angle, spring deflection, external joint torque signal, and target motor torque; A processing module is used to process the linear spring deformation length and the deflection angle based on a preset neural network model to generate a lower limb exoskeleton static interaction force and an estimated motor torque; process the external joint torque signal based on a preset neural network model to generate an excitation torque; process the exoskeleton trunk segment position, the exoskeleton trunk segment direction, the exoskeleton right hip joint, the exoskeleton knee joint angle, the exoskeleton left hip joint, and the exoskeleton knee joint angle based on a preset neural network model to generate a lower limb exoskeleton passive power value; process the excitation torque, the lower limb exoskeleton passive power value and the lower limb exoskeleton static interaction force based on a preset neural network model to generate an exoskeleton measurement value; process the target motor torque and the estimated motor torque to generate a motor torque estimation error value; process the preset neural network model based on the motor torque estimation error value to generate a lower limb exoskeleton dynamic interaction force.

9. An electronic device, characterized in that: include: a first processor; and a memory for storing executable instructions of the first processor; Wherein, the first processor is configured to execute the method for measuring the interaction force of the lower limb exoskeleton as described in any one of claims 1 to 7 by executing the executable instructions.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the second processor, the method for measuring the interaction force of the lower limb exoskeleton as claimed in any one of claims 1 to 7 is implemented.

Images