Soft grasping control method based on myoelectricity decoding grasping force and muscle rigidity

Abstract

The invention discloses a flexible grasping control method based on myoelectricity decoding grasping force and rigidity, and aims to solve the problems that a traditional prosthetic hand only considers the grasping force, the prosthetic hand applies the grasping force in a fixed and rigid movement mode, and an object is easily damaged. According to the method, the gripping force of the human hand is estimated by adopting a biological muscle model, meanwhile, the muscle stiffness of the human hand is estimated in combination with the contraction characteristics of the muscles, and the mapping relation between the muscle stiffness of the human hand and the stiffness parameter of the prosthetic hand controller is established according to the characteristics when the human hand grips an object; and the flexible grasping control of the prosthetic hand on different objects is realized. According to the artificial limb hand, the hand gripping force can be obtained, the muscle rigidity of the hand in the object gripping process can be further obtained, the object is gripped according to the expected gripping force and gripping rigidity, compared with a transmission method, the natural control feeling more similar to that of an original limb can be provided, and the use experience of the artificial limb hand is improved.

Description

Technical Field

[0001] The present invention relates to a compliant grasping control method based on electromyographic decoding of grasping force and muscle stiffness. The method controls a prosthetic hand by performing electromyographic decoding on the grasping force and muscle stiffness of a human hand during grasping based on a biological muscle model. The method can actively adjust the desired force and stiffness parameters of a controller according to the grasping intention of the human hand, thereby realizing compliant grasping control based on the grasping intention of the human hand. Background Art

[0002] Myoelectrically controlled prosthetic hands are an advanced technology that uses electromyographic signals to drive and control the movement of prosthetic hands. When a user tries to move the muscles on their residual limbs, these muscles generate weak electrical signals. By placing electrodes on the surface of the skin or inside the muscles, these signals can be captured and transmitted to the control system of the prosthetic hand. After signal amplification, filtering and processing, the control system converts the user's muscle activity into corresponding prosthetic movements, such as opening and closing, grasping and rotating the hand. In this way, the prosthetic hand can respond to the user's intentions more naturally and flexibly, allowing the user to experience higher control accuracy and reliability when using the prosthetic hand. The application of this technology not only significantly improves the functionality of prostheses, but also greatly improves the quality of life of users. Myoelectrically controlled prosthetic hands can help users complete various basic operations in daily life, such as picking up objects, opening and closing doors and windows, etc., so that they can live more independently. In addition, this technology can also be used for more complex tasks, such as operating tools and performing delicate manual operations, showing its great potential in the fields of rehabilitation medicine and assistive technology.

[0003] However, the current grasping decoding of human hands mainly focuses on the decoding of grasping force. The grasping action of human hands includes not only grasping force but also muscle stiffness. In order to avoid damage to objects, human hands often use smaller muscle stiffness to touch objects. Therefore, in order to achieve compliant grasping of objects, on the basis of estimating the grasping force, it is necessary to further estimate the muscle stiffness during grasping. On the other hand, in order to enable the prosthetic hand to better simulate the human hand and avoid damage to the grasped object, it is necessary to control the prosthetic hand to perform compliant grasping based on the grasping force and muscle stiffness obtained by the electromyographic signal. Considering that impedance control can show the desired response characteristics when the robot's motion is affected by the environment by designing virtual impedance mass stiffness damping. By adjusting the stiffness and damping parameters, the system can show compliant characteristics to adapt to complex environments, and can be used for compliant grasping control based on electromyographic signals. However, the parameter selection in the impedance controller depends on engineering experience, and the adaptability is poor when contacting unknown environments, which is prone to large control errors and instability. Therefore, obtaining the grasping force and muscle stiffness from the electromyographic signals and combining them with the compliant control method to comprehensively adjust the expected force and stiffness parameters of the prosthetic hand controller can better reflect the grasping intention of the human hand and further expand the usage scenarios of the prosthetic hand. Summary of the invention

[0004] The purpose of the present invention is to propose a compliant grasping control method based on myoelectric decoding of grasping force and muscle stiffness, by establishing a dynamic biological model of muscle contraction to decode the grasping force and muscle stiffness information in the myoelectric signal, and controlling the prosthetic hand according to the grasping force and muscle stiffness obtained, so as to achieve compliant grasping control of the prosthetic hand. Firstly, a grasping force estimation method of a human hand is established based on the Hulexy biological muscle model, and then a muscle stiffness estimation method of a human hand is further established according to the muscle contraction principle, and then a mapping relationship between the muscle stiffness of a human hand and the stiffness parameter of a prosthetic hand controller is established based on a fuzzy logic relationship, and finally a compliant grasping control method of an underactuated prosthetic hand is proposed in combination with an impedance control algorithm.

[0005] The technical solution adopted by the present invention to solve its technical problem is:

[0006] A compliant grasping control method based on myoelectric decoding of grasping force and muscle stiffness comprises the following steps:

[0007] Step 1: Establish a biological muscle contraction model when the prosthetic hand grasps an object, and estimate the grasping force and muscle stiffness based on the electromyographic signal based on the model;

[0008] Step 2: Establish the kinematic and dynamic models of the prosthetic hand according to the structure and motion relationship of the prosthetic hand;

[0009] Step 3: According to the operation characteristics of human hands, a mapping relationship between the muscle stiffness of human hands when grasping objects and the stiffness parameters of the prosthetic hand controller is established;

[0010] Step 4: Combining the hand grip force and muscle stiffness decoded from the electromyographic signal, a compliant control method for grip force and muscle stiffness based on electromyographic decoding is proposed. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 It is the overall algorithm structure and logic diagram of the present invention.

[0012] Figure 2 It is the human hand grasping force estimated by the electromyographic signal in the present invention.

[0013] Figure 3 It is the muscle stiffness of the human hand estimated by the electromyographic signal in the present invention.

[0014] Figure 4 It is a schematic structural diagram of a finger of an implementation object of the present invention.

[0015] Figure 5 This is a block diagram of the compliant grasping control algorithm of the present invention.

[0016] Figure 6 It is the contact force when the embodiment of the present invention grasps an object.

[0017] Figure 7 It is the driving voltage when the object of the present invention grasps an object. DETAILED DESCRIPTION

[0018] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments.

[0019] The specific steps of a compliant grasping control method based on myoelectric decoding of grasping force and muscle stiffness are as follows:

[0020] Step 1: Establish a biological muscle contraction model when the prosthetic hand grasps an object, and estimate the grasping force and muscle stiffness based on the electromyographic signal according to the model; the model is a nonlinear partial differential equation, and the model parameters in the model are related to the biological characteristics of muscle contraction:

[0021]

[0022] Where v(t) represents the slip rate of the myofilament, x represents the distance between the cross-bridge binding site and the myosin head, p(x, t) is the distribution function of the number of cross-bridge bindings, f(x, t) represents the separation rate function, g(x, t) represents the binding rate function, and r(t) represents the muscle activation function related to the electromyographic signal;

[0023] Secondly, the balanced truncation method is used to further reduce the dimension and obtain an extremely low-dimensional skeletal muscle dynamics model; finally, the time variable is integrated and a low-dimensional muscle mechanics model is obtained based on the relationship between force and speed:

[0024]

[0025] in is the variable after the model dimension reduction, is the time-varying system coefficient after dimension reduction, is the correlation coefficient of electromyographic signal, C k is the muscle force correlation coefficient;

[0026] Considering the muscle contraction length within time T, we have:

[0027]

[0028] Since muscle contraction speed is usually approximately related to grip strength:

[0029] (F+F a )(v(t)+v b )=(F0+F a )v b (5)

[0030] Where F0 is the maximum isometric contraction force of the muscle F a and v b They are fixed coefficients;

[0031] After substitution:

[0032]

[0033] According to the above formula, the stiffness of the muscle in this interval can be calculated as:

[0034]

[0035] The results of grip force and muscle stiffness decoded from electromyographic signals when the human hand grasps three different objects are shown in Figure 2. Figure 2 and Figure 3 As shown; it can be seen that the plastic cup (object 1) has the greatest stiffness and is not easily deformed or damaged; therefore, when grasping the plastic cup (object 1), the human hand uses a larger gripping force and muscle stiffness to grasp it; the single-layer paper cup (object 2) and the double-layer paper cup (object 3) are prone to over-deformation and damage, so when grasping such objects, the human hand uses a relatively smaller gripping force and muscle stiffness to grasp the objects.

[0036] Step 2: Figure 4 The structure of the prosthetic hand shown in the figure establishes the kinematic and dynamic models of the prosthetic hand. In this example, the Lagrangian method is used to establish the dynamic equations of the thumb hand:

[0037]

[0038]

[0039] Where, represents the moment of inertia of the nth connecting rod I n , m n represents the mass of the nth connecting rod, represents the coordinates of the center of mass of the nth link, g represents the acceleration of gravity, τ1 represents the driving force of link 1, F represents the grip force, and f represents the friction force;

[0040] Finally, the fingertip contact dynamics model of the prosthetic hand can be obtained as follows:

[0041]

[0042] Where θ is the rotation angle of the prosthetic finger; M is the mass matrix, C is the velocity related term, G is the gravity matrix, τ is the joint driving force, J is the Jacobian matrix, and F is the resultant external force;

[0043] Step 3, according to the operation characteristics of human hands, a mapping relationship between the muscle stiffness of human hands when grasping objects and the stiffness parameters of the prosthetic hand controller is established;

[0044] Use Figure 5 The control method shown in the figure can provide a natural control feeling similar to the original limb when the impedance parameter is set to a value similar to that of the human hand. When the muscle stiffness of the human hand is large, the stiffness parameter of the controller should be increased accordingly. In order to realize the anthropomorphic muscle stiffness adjustment strategy, the following fuzzy rules are used to adjust the stiffness of the object.

[0045] When the muscle stiffness and its variation are large, the stiffness parameter of the controller is higher to ensure that the prosthetic hand can grasp the object quickly and stably. When the muscle stiffness and its variation are small, the stiffness parameter of the controller is lower to enable the prosthetic hand to grasp the object smoothly when in contact with it.

[0046] Step 4, combining the hand grip force and muscle stiffness decoded from the electromyographic signal to propose a compliant control method for grip force and muscle stiffness based on electromyographic decoding;

[0047] In order to achieve a smooth grasping action, the magnitude of the grasping force and the convergence rate of the force should be adjusted according to the characteristics of the grasped object; to avoid the prosthetic hand from damaging the object, the grasping force of the prosthetic hand should be small and slow to achieve a slow and gentle grasp; for objects that are not easily damaged, it is expected that the grasping force can converge quickly to achieve a stable grasp;

[0048] When the prosthetic hand grasps an object, in order to simulate the muscle stiffness of the human hand, the following controller can be designed based on impedance control:

[0049]

[0050] where θ d is the desired rotation angle, M d is the expected quality parameter, B d is the desired damping parameter, K d is the expected stiffness parameter, F e is the grip force error, e is the position error

[0051] In order to verify the compliant grasping control method of the underactuated myoelectric prosthetic hand proposed in this paper, grasping experiments of different types of objects were designed for verification. During the experiment, the prosthetic hand was used to grasp three objects: plastic cups, single-layer paper cups and double-layer paper cups. Among them, soft plastic cups and soft paper cups are prone to excessive deformation during grasping, resulting in irreversible damage to the grasped objects. Plastic cups were used as the control group and were not easily damaged during the grasping process. The grasping experiments on the above three objects were used to verify the grasping ability of the compliant grasping control method based on muscle stiffness estimation proposed in this paper for different objects. The experimental results are shown in Figure 2. Figure 6 and Figure 7As shown in the figure, the operation strategy of the human hand is reflected in the prosthetic hand through the electromyographic signal. When grasping the plastic cup (object 1), the prosthetic hand adopts a larger expected force and stiffness parameter to achieve a fast and stable grasp; when grasping the single-layer paper cup (object 2) and the double-layer paper cup (object 3), the prosthetic hand also adopts a smaller expected force and stiffness parameter to achieve a slow and gentle grasp to avoid damage to the object.

[0052] Traditional prosthetic hands only consider the grasping force, and all prosthetic hands use fixed and rigid movements to apply grasping force. The method proposed in this paper can further obtain the muscle stiffness of the human hand when grasping an object, and grasp the object according to the desired grasping force and grasping stiffness. Compared with the transmission method, it can provide a more natural control feeling similar to the original limb. Therefore, this method has good research and application prospects in the grasping of prosthetic hands. It can achieve human-like grasping without adding additional sensors, and can provide a new anthropomorphic grasping control method for low-cost prosthetic hands.

Claims

1. A method for controlling compliant grasping based on myoelectric decoding of grasping force and muscle stiffness, which is suitable for controlling a myoelectric prosthetic hand to compliantly grasp an object. The method decodes the grasping force and muscle stiffness of a human hand grasping an object using myoelectric signals. By establishing a mapping relationship between the muscle stiffness of a human hand and the stiffness parameters of a prosthetic hand controller, the method can achieve compliant grasping of objects with different characteristics by a prosthetic hand. The method is characterized in that: Firstly, a method for estimating the grip force of human hands using electromyographic signals is established based on the biological muscle model. Secondly, a method for estimating muscle stiffness using electromyographic signals is established in combination with the muscle contraction principle. Then, according to the operation characteristics of the human hand, the mapping relationship between the muscle stiffness of the human hand and the stiffness parameter of the controller is established; finally, the compliant grasping control based on the myoelectric decoding grasping force and muscle stiffness is realized in combination with the model of the controlled object; the specific steps of the compliant grasping control method based on the myoelectric decoding grasping force and stiffness are as follows: The first step is to establish a method for estimating the grip force of human hands using electromyographic signals based on the biological muscle model: Where v(t) represents the slip rate of the myofilament, x represents the distance between the cross-bridge binding site and the myosin head, p(x, t) is the cross-bridge binding number distribution function; f(x, t) represents the separation rate function; g(x, t) represents the binding rate function, and r(t) represents the muscle activation function related to the electromyographic signal; Then, the balanced truncation method is used to further reduce the dimension and obtain an extremely low-dimensional skeletal muscle dynamics model; finally, the time variable is integrated and a low-dimensional muscle mechanics model is obtained based on the relationship between force and speed: in is the variable after the model dimension reduction, is the time-varying system coefficient after dimension reduction, is the correlation coefficient of electromyographic signal, C k is the muscle force correlation coefficient; The second step is to establish a method for estimating muscle stiffness using electromyographic signals based on the muscle contraction principle. Considering the muscle contraction length within time T, we have: Since muscle contraction speed is usually approximately related to grip strength: (F+F a )(v(t)+v b )=(F0+F a )v b (5) Where F0 is the maximum isometric contraction force of the muscle F a and v b They are fixed coefficients; After substitution: According to the above formula, the stiffness of the muscle in this interval can be calculated as: The third step is to establish a mapping relationship between the muscle stiffness of the human hand and the stiffness parameters of the controller according to the operation characteristics of the human hand; When the impedance parameter is set to a value similar to that of a human hand, it can provide a natural sense of control similar to that of the original limb. When the muscle stiffness of the human hand is large, the stiffness parameter of the controller should be increased accordingly. In order to implement an anthropomorphic muscle stiffness adjustment strategy, the following fuzzy rules are used to adjust the stiffness of the object. When the muscle stiffness and its variation are large, the stiffness parameter of the controller is high to ensure that the prosthetic hand can grasp the object quickly and stably. When the muscle stiffness and its variation are small, the stiffness parameter of the controller is low to enable the prosthetic hand to grasp the object smoothly when in contact with it. The fourth step is to realize compliant grasping control based on myoelectric decoding grasping force and muscle stiffness in combination with the model of the controlled object; Consider a kinetic model of the following form: Where θ is the rotation angle of the prosthetic finger; M is the mass matrix, C is the velocity related term, G is the gravity matrix, τ is the joint driving force, J is the Jacobian matrix, and F is the resultant external force; Design the following controller based on impedance control: where θ d is the desired rotation angle, M d is the expected quality parameter, B d is the desired damping parameter, K d is the expected stiffness parameter, F e is the gripping force error, and e is the position error.

2. According to the mapping relationship described in the third step of claim 1, it is characterized in that The estimated muscle stiffness of the human hand and the stiffness parameter of the controller can be mapped to values ​​with similar trends.

3. The muscle activation function according to the first step of claim 1, characterized in that It can express the relationship between EMG signals and muscle activation degree, including proportional function, exponential function and polynomial function.

Images