A method, device, and processor for amplifying lower limb exoskeleton sensitivity control based on adaptive parameter estimation.

Abstract

The present invention discloses a sensitivity amplification control method for lower limb exoskeleton based on adaptive parameter estimation, including establishing a two-degree-of-freedom dynamic model of a single leg of a lower limb exoskeleton robot; online estimating unknown parameters of the established two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot to obtain an estimated two-degree-of-freedom dynamic model of the single leg of the lower limb exoskeleton robot; substituting the estimated two-degree-of-freedom dynamic model of the single leg of the lower limb exoskeleton robot into a sensitivity amplification controller to obtain a required control law. Aiming at the problems existing in sensitivity amplification control, the present invention proposes an online parameter estimation method for designing the parameters required by the controller. The improved sensitivity amplification control ensures that the parameters converge to the true value while also achieving a high degree of overlap between the joint trajectory of the exoskeleton and the walking trajectory of the human body.

Description

Technical Field

[0001] The invention relates to a lower limb exoskeleton sensitivity amplification control method based on adaptive parameter estimation, belonging to the technical field of exoskeleton control. Background Art

[0002] The lower limb exoskeleton robot is an anthropomorphic bionic structure that can be worn on the human body. It can meet the human body's gait needs, provide assistance to the human body, enhance human functions, and complete certain functions and tasks under the control of the operator. It has very broad application prospects both in the military and civilian fields.

[0003] The application of advanced control methods can bring better human-machine interaction performance to exoskeleton robots. At present, many control methods have been applied to lower limb exoskeleton robot systems, such as fuzzy control, sliding mode control, anti-disturbance control, adaptive impedance control, etc., and sensitivity amplification control has been successfully applied to exoskeletons such as BLEEX and HULC as early as the beginning of the 21st century due to its excellent performance. The sensitivity amplification control algorithm can ensure that the interaction force between the exoskeleton and the wearer remains in a small area of ​​0. In addition, this control method does not require direct measurement of human-machine interaction forces, so it can reduce the complexity of the exoskeleton system. In classical and modern control theories, it is usually necessary to minimize the sensitivity of the system to external disturbances. But for exoskeleton control, a completely opposite goal is required: to maximize the sensitivity of the closed-loop system to external forces. However, the high sensitivity of the system to external forces causes the system to lose robustness. This requires improving the accuracy of the exoskeleton dynamics model. Summary of the invention

[0004] The invention provides a lower limb exoskeleton sensitivity amplification control method based on adaptive parameter estimation, which is used for controlling a lower limb exoskeleton robot.

[0005] The technical solution of the present invention is: a lower limb exoskeleton sensitivity amplification control method based on adaptive parameter estimation, comprising:

[0006] Establish a two-degree-of-freedom dynamic model of a single leg of a lower limb exoskeleton robot;

[0007] The unknown parameters of the established two-degree-of-freedom dynamic model of the single leg of the lower limb exoskeleton robot are estimated online to obtain the estimated two-degree-of-freedom dynamic model of the single leg of the lower limb exoskeleton robot;

[0008] The estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot is substituted into the sensitivity amplification controller to obtain the required control law.

[0009] The method of establishing a two-degree-of-freedom dynamic model of a single leg of a lower limb exoskeleton robot comprises: using a lower limb exoskeleton geometric model to obtain an expression for the distance between a rod and a rotation center, and an expression for the velocity of the rod's center of mass, and determining the system kinetic energy and the system potential energy based on the expressions, and finally obtaining a two-degree-of-freedom dynamic model of a single leg of a lower limb exoskeleton robot.

[0010] The two-degree-of-freedom dynamic model of a single leg of a lower limb exoskeleton robot is established, comprising:

[0011] The Lagrangian L is defined as:

[0012] L=E k -E P (1)

[0013] Where: E k is the kinetic energy of the lower limb exoskeleton robot system, E p The potential energy of the lower limb exoskeleton robot system;

[0014] The Lagrangian equation is obtained by taking the partial derivative of the Lagrangian L with respect to q and then differentiating it with respect to time t:

[0015]

[0016] Where: T is the generalized external force vector in motion, q is the generalized coordinate vector of the system;

[0017] According to the geometric model of the lower limb exoskeleton, the distance from the mass center 12 of the thigh rod to the rotation center 10 of the hip joint and the distance from the mass center 13 of the calf rod to the rotation center 11 of the knee joint are obtained as follows:

[0018]

[0019] Where: l c,h is the distance from the center of mass 12 of the thigh bar to the center of rotation 10 of the hip joint, l c,k is the distance from the mass center 13 of the shank bar to the rotation center 11 of the knee joint, x c,h y c,h is the spatial coordinate of the mass center 12 of the thigh member, x c,k y c,k is the spatial coordinate of the mass center 13 of the calf rod;

[0020] The velocity of the bar's center of mass is:

[0021]

[0022] Where: v h is the velocity of the mass center 12 of the thigh member, v k is the velocity of the mass center 13 of the shank member, is the angular velocity of the hip joint rotation center 10, is the angular velocity of the knee joint rotation center 11;

[0023] The lower limb exoskeleton robot has two active degrees of freedom per leg, and the kinetic energy of the lower limb exoskeleton robot system is:

[0024]

[0025] Where: E k,h 、E k,k are the kinetic energy of the thigh rod 6 and the calf rod 8 respectively; m h 、m k I is the mass of the hip joint rotation center 10 of the thigh rod 6 and the calf rod 8 respectively; h ,I k are the moments of inertia of the thigh rod 6 and the calf rod 8 respectively; Angular acceleration of the hip joint rotation center 10 and the knee joint rotation center 11 respectively; q h , are the angle, angular velocity and angular acceleration of the hip joint rotation center 10 respectively; q k , are the angle, angular velocity, and angular acceleration of the knee joint rotation center 11, respectively;

[0026] The potential energy of the lower limb exoskeleton robot system is:

[0027] E p =m h gl c,h cosq h +m k g(l c,h cosq h +l c,k cosq k )(6)

[0028] Where: g is the acceleration due to gravity;

[0029] Based on the Lagrange equation, the two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot is obtained:

[0030]

[0031] Where: M(q) is the inertia matrix, is the centripetal force matrix, G(q) is the gravity vector; τ is the driving torque generated by the driver of the lower limb exoskeleton robot system.

[0032] The inertia matrix, centripetal force matrix, gravity vector, and driving torque of the lower limb exoskeleton system are as follows:

[0033]

[0034] Where: M 11 (q), M 12 (q), M 21 (q), M 22 (q) represents the element in the inertia matrix; represents the elements in the centripetal force matrix, G1(q) and G2(q) represent the elements in the gravity vector.

[0035] The method comprises: performing online estimation on the unknown parameters of the established two-degree-of-freedom dynamic model of a single leg of a lower limb exoskeleton robot to obtain an estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot, comprising: performing linear parameterization processing on the obtained two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot, introducing a first-order low-pass filtering operation, and designing an adaptive law driven by parameter estimation errors to effectively estimate the unknown parameters.

[0036] The method of online estimating unknown parameters of the established two-degree-of-freedom dynamics model of a single leg of the lower limb exoskeleton robot to obtain an estimated two-degree-of-freedom dynamics model of a single leg of the lower limb exoskeleton robot comprises:

[0037] Rewrite the left side of the two-degree-of-freedom dynamic model equation of the single leg of the lower limb exoskeleton robot into a linear parameterized form:

[0038]

[0039] Where: M(q) is the inertia matrix, is the centripetal force matrix, G(q) is the gravity vector, the inertia matrix, the centripetal force matrix, and the gravity vector are matrices containing unknown parameters in the two-degree-of-freedom dynamics model of the single leg of the lower limb exoskeleton robot; θ is a matrix about m h ,m k The constant vector to be estimated, is the known regression matrix;

[0040] Then define the intermediate parameters

[0041] get:

[0042]

[0043] Where: Does not include acceleration terms The regression matrix of

[0044] The two-degree-of-freedom dynamics model of a single leg of the lower limb exoskeleton robot is rewritten to obtain a reconstructed two-degree-of-freedom dynamics model of a single leg of the lower limb exoskeleton robot:

[0045]

[0046] Where:

[0047] Low-pass filtering is introduced on both sides of the two-degree-of-freedom dynamic model equation of the reconstructed lower limb exoskeleton robot single leg to obtain:

[0048]

[0049] Where: They are The regression matrix after filtering; τ f represents τ after low-pass filtering;

[0050]

[0051] Where: k represents the low-pass filter coefficient, t represents time;

[0052] Substituting formula (12) into formula (11), we obtain:

[0053]

[0054] Where: is a new regression matrix that does not contain acceleration

[0055] Then define the matrix P1 and vector Q1:

[0056]

[0057] Where: l>0 is an adjustable parameter; express The transpose of ; P1(0) and Q1(0) represent the values ​​of matrix P1 and vector Q1 at the initial time t=0 respectively;

[0058] Define vector W1 as:

[0059]

[0060] Where: express The error with θ;

[0061] Design the corresponding system estimation parameters The adaptive law is:

[0062]

[0063] Where: Γ>0 is the learning gain matrix;

[0064] By estimating Substituting the value back into formula (7), we can obtain the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot:

[0065]

[0066] Where: They respectively represent the inertia matrix, centripetal force matrix, and gravity vector of the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot.

[0067] Substituting the estimated two-degree-of-freedom dynamics model of a single leg of the lower limb exoskeleton robot into the sensitivity amplification controller to obtain the required control law includes:

[0068] Definition G hm is a model of human-computer interaction, G exo is the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot, τ h is the human-machine interaction force, τ is the driving torque generated by the actuator of the lower limb exoskeleton robot system, Represents the output trajectory parameters of the lower limb exoskeleton robot system, then:

[0069] q = τ h G hm +τG exo (18)

[0070] τ s =τ+τ h (19)

[0071] Where: τ s It represents the torque generated by the joints in the lower limb exoskeleton robot system under the human-robot interaction model;

[0072] The designed sensitivity amplification control law is:

[0073]

[0074] Where: α is the sensitivity amplification factor; They respectively represent the inertia matrix, centripetal force matrix, and gravity vector of the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot.

[0075] The lower limb exoskeleton geometric model is simplified by the lower limb exoskeleton robot system: the hip joint pulley module 5 is simplified to the hip joint rotation center 10, the knee joint pulley module 7 is simplified to the knee joint rotation center 11, the thigh rod mass center 12 is used as the thigh rod 6 mass center, and the calf rod mass center 13 is used as the calf rod 8 mass center.

[0076] According to another aspect of an embodiment of the present invention, there is provided a lower limb exoskeleton sensitivity amplification control device based on adaptive parameter estimation, comprising:

[0077] The first module is used to establish a two-degree-of-freedom dynamic model of a single leg of a lower limb exoskeleton robot;

[0078] The second module is used to perform online estimation of unknown parameters of the established two-degree-of-freedom dynamics model of a single leg of the lower limb exoskeleton robot, and obtain an estimated two-degree-of-freedom dynamics model of the single leg of the lower limb exoskeleton robot;

[0079] The third module is used to substitute the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot into the sensitivity amplification controller to obtain the required control law.

[0080] According to another aspect of an embodiment of the present invention, a processor is provided, wherein the processor is used to run a program, wherein when the program is run, any one of the above-mentioned lower limb exoskeleton sensitivity amplification control methods based on adaptive parameter estimation is executed.

[0081] The beneficial effects of the present invention are:

[0082] 1. Compared with other model-based exoskeleton control algorithms, the sensitivity amplification control algorithm does not require direct measurement of human-computer interaction force, which can reduce the complexity of the exoskeleton system. On this basis, the control method provided by the present invention further ensures that the system is highly sensitive to external forces without losing the robustness of the system.

[0083] 2. Since the sensitivity amplification control algorithm is a model-based control algorithm, the accuracy of its dynamic model directly affects the control effect of the algorithm. The adaptive parameter estimation algorithm provided by the present invention can accurately estimate the parameters of the lower limb exoskeleton dynamic model online, thereby improving the performance of the sensitivity amplification control algorithm.

[0084] In summary, in order to address the problems existing in sensitivity amplification control, the present invention proposes an online parameter estimation method for designing the parameters required by the controller. The improved sensitivity amplification control not only ensures that the parameters converge to the true values, but also achieves a high degree of overlap between the joint trajectory of the exoskeleton and the walking trajectory of the human body. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Figure 1 This is a block diagram of the sensitivity amplification control principle of the present invention;

[0086] Figure 2 This is a diagram of the mechanical structure of a single-leg lower limb exoskeleton;

[0087] Figure 3 It is the geometric model of lower limb exoskeleton;

[0088] Figure 4 Tracking graph for hip angle;

[0089] Figure 5 It is a curve graph for tracking knee angle;

[0090] Figure 6 This is the effect diagram of adaptive parameter estimation;

[0091] The numbers in the figure are: 1-winding wheel, 2-power system, 3-pulley module, 4-waist component, 5-hip joint pulley module, 6-thigh rod, 7-knee joint pulley module, 8-calf rod, 9-human trunk, 10-hip joint rotation center, 11-knee joint rotation center, 12-thigh rod center of mass, 13-calf rod center of mass. DETAILED DESCRIPTION

[0092] The invention is further described below in conjunction with the accompanying drawings and embodiments, but the content of the invention is not limited to the scope of the embodiments.

[0093] Example 1: Figure 1-6 As shown, a sensitivity amplification control method of a lower limb exoskeleton based on adaptive parameter estimation is used to establish a two-degree-of-freedom dynamic model of a single leg of a lower limb exoskeleton robot;

[0094] The unknown parameters of the established two-degree-of-freedom dynamic model of the single leg of the lower limb exoskeleton robot are estimated online to obtain the estimated two-degree-of-freedom dynamic model of the single leg of the lower limb exoskeleton robot;

[0095] The estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot is substituted into the sensitivity amplification controller to obtain the required control law.

[0096] Furthermore, the two-degree-of-freedom dynamic model of a single leg of a lower limb exoskeleton robot can be set up, including: using the lower limb exoskeleton geometric model to obtain an expression for the distance between the rod and the center of rotation, and an expression for the velocity of the center of mass of the rod, and determining the system kinetic energy and system potential energy based on the expressions, and finally obtaining a two-degree-of-freedom dynamic model of a single leg of a lower limb exoskeleton robot.

[0097] Furthermore, the two-degree-of-freedom dynamic model of a single leg of a lower limb exoskeleton robot may be set to include:

[0098] The Lagrangian L is defined as:

[0099] L=E k -E P (1)

[0100] Where: E k is the kinetic energy of the lower limb exoskeleton robot system, E p The potential energy of the lower limb exoskeleton robot system;

[0101] The Lagrangian equation is obtained by taking the partial derivative of the Lagrangian L with respect to q and then differentiating it with respect to time t:

[0102]

[0103] Where: T is the generalized external force vector (including force and torque) in motion, q is the generalized coordinate vector of the system (the Lagrange equation defined here is a universal equation. In the subsequent model, it represents the output trajectory parameter of the lower limb exoskeleton robot system, corresponding to q h ,q k );

[0104] according to Figure 3 The geometric model of the lower limb exoskeleton is as follows: The distance from the center of mass 12 of the thigh bar to the center of rotation 10 of the hip joint, and the distance from the center of mass 13 of the calf bar to the center of rotation 11 of the knee joint are obtained as follows:

[0105]

[0106] Where: l c,h is the distance from the center of mass 12 of the thigh bar to the center of rotation 10 of the hip joint, l c,k is the distance from the mass center 13 of the shank bar to the rotation center 11 of the knee joint, x c,h y c,h is the spatial coordinate of the mass center 12 of the thigh member, x c,k y c,k is the spatial coordinate of the mass center 13 of the calf rod;

[0107] The velocity of the bar's center of mass is:

[0108]

[0109] Where: v h is the velocity of the mass center 12 of the thigh member, v k is the velocity of the mass center 13 of the shank member, is the angular velocity of the hip joint rotation center 10, is the angular velocity of the knee joint rotation center 11;

[0110] The lower limb exoskeleton robot has two active degrees of freedom per leg (including the hip joint and the knee joint). The kinetic energy of the lower limb exoskeleton robot system is:

[0111]

[0112] Where: E k,h 、E k,k are the kinetic energy of the thigh rod 6 and the calf rod 8 respectively; m h 、m kI is the mass of the hip joint rotation center 10 of the thigh rod 6 and the calf rod 8 respectively; h ,I k are the moments of inertia of the thigh rod 6 and the calf rod 8 respectively; Angular acceleration of the hip joint rotation center 10 and the knee joint rotation center 11 respectively; q h , are the angle, angular velocity and angular acceleration of the hip joint rotation center 10 respectively; q k , are the angle, angular velocity, and angular acceleration of the knee joint rotation center 11, respectively;

[0113] The potential energy of the lower limb exoskeleton robot system is:

[0114] E p =m h gl c,h cosq h +m k g(l c,h cosq h +l c,k cosq k )(6)

[0115] Where: g is the acceleration due to gravity;

[0116] Based on the Lagrange equation, the two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot is obtained:

[0117]

[0118] Where: M(q) is the inertia matrix, is the centripetal force matrix, G(q) is the gravity vector; τ is the driving torque generated by the driver of the lower limb exoskeleton robot system.

[0119] Furthermore, the inertia matrix, the centripetal force matrix, the gravity vector, and the driving torque of the lower limb exoskeleton system may be set as follows:

[0120]

[0121] Where: M 11 (q), M 12 (q), M 21 (q), M 22 (q) represents the element in the inertia matrix; represents the elements in the centripetal force matrix, G1(q) and G2(q) represent the elements in the gravity vector.

[0122] Furthermore, the unknown parameters of the established two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot can be set to be estimated online to obtain an estimated two-degree-of-freedom dynamic model of the single leg of the lower limb exoskeleton robot, including: linear parameterization of the obtained two-degree-of-freedom dynamic model of the single leg of the lower limb exoskeleton robot, eliminating the joint angular acceleration term by introducing a first-order low-pass filtering operation, and designing an adaptive law driven by parameter estimation errors to effectively estimate the unknown parameters.

[0123] Specifically, the matrix containing unknown parameters in the two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot is: M(q), G(q), the obtained two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot is linearly parameterized, and the joint angular acceleration term is eliminated by introducing a first-order low-pass filtering operation. Under the condition of continuous excitation, an adaptive law driven by parameter estimation error is designed to effectively estimate the unknown parameters.

[0124] Furthermore, the unknown parameters of the established two-degree-of-freedom dynamics model of the single leg of the lower limb exoskeleton robot can be estimated online to obtain the estimated two-degree-of-freedom dynamics model of the single leg of the lower limb exoskeleton robot, including:

[0125] Rewrite the left side of the two-degree-of-freedom dynamic model equation of the single leg of the lower limb exoskeleton robot into a linear parameterized form:

[0126]

[0127] Where: M(q) is the inertia matrix, is the centripetal force matrix, G(q) is the gravity vector, the inertia matrix, the centripetal force matrix, and the gravity vector are matrices containing unknown parameters in the two-degree-of-freedom dynamics model of the single leg of the lower limb exoskeleton robot; θ is a matrix about m h ,m k The constant vector to be estimated, is the known regression matrix;

[0128] Then define the intermediate parameters

[0129] get:

[0130]

[0131] Where: Does not include acceleration terms The regression matrix of

[0132] The two-degree-of-freedom dynamics model of a single leg of the lower limb exoskeleton robot is rewritten to obtain a reconstructed two-degree-of-freedom dynamics model of a single leg of the lower limb exoskeleton robot:

[0133]

[0134] Where:

[0135] Low-pass filtering is introduced on both sides of the two-degree-of-freedom dynamic model equation of the reconstructed lower limb exoskeleton robot. get:

[0136]

[0137] Where: They are The regression matrix after filtering; τ f represents τ after low-pass filtering;

[0138]

[0139] Where: k represents the low-pass filter coefficient, t represents time;

[0140] Substituting formula (12) into formula (11), we obtain:

[0141]

[0142] Where: is a new regression matrix that does not contain acceleration

[0143] Then define the matrix P1 and vector Q1:

[0144]

[0145] Where: e>0 is an adjustable parameter; express The transpose of ; P1(0) and Q1(0) represent the values ​​of matrix P1 and vector Q1 at the initial time t=0 respectively;

[0146] Define vector W1 as:

[0147]

[0148] Where: express The error with θ;

[0149] Design the corresponding system estimation parameters The adaptive law is:

[0150]

[0151] Where: Γ>0 is the learning gain matrix;

[0152] By estimating Substituting the value back into formula (7), we can obtain the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot:

[0153]

[0154] Where: They respectively represent the inertia matrix, centripetal force matrix, and gravity vector of the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot.

[0155] Furthermore, the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot may be substituted into the sensitivity amplification controller to obtain the required control law, including:

[0156] like Figure 1 As shown, define G hm is a model of human-computer interaction, G exo is the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot, G e ' xo is the estimated two-degree-of-freedom inverse dynamics model of a single leg of the lower limb exoskeleton robot, τ h is the human-machine interaction force, τ is the driving torque generated by the actuator of the lower limb exoskeleton robot system, Represents the output trajectory parameters of the lower limb exoskeleton robot system, then:

[0157] q = τ h G hm +τG exo (18)

[0158] τ s =τ+τ h (19)

[0159] Where: τ s It represents the torque generated by the joints in the lower limb exoskeleton robot system under the human-robot interaction model;

[0160] The designed sensitivity amplification control law is:

[0161]

[0162] Where: α is the sensitivity amplification factor; They respectively represent the inertia matrix, centripetal force matrix, and gravity vector of the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot.

[0163] By using formula (7) and formula (20), we can get:

[0164]

[0165] Furthermore, the lower limb exoskeleton robot system is simplified into a lower limb exoskeleton geometric model: the hip joint pulley module 5 is simplified to the hip joint rotation center 10, the knee joint pulley module 7 is simplified to the knee joint rotation center 11, the thigh rod center of mass 12 is used as the center of mass of the thigh rod 6, and the calf rod center of mass 13 is used as the center of mass of the calf rod 8.

[0166] Specifically, the lower limb exoskeleton robot system is Figure 2 As shown, it includes: a winding wheel 1, a power system 2, a pulley module 3, a waist component 4, a thigh rod 6, and a calf rod 8; the pulley module 3 includes a hip joint pulley module 5 and a knee joint pulley module 7, the power system 2 installed on the waist component 4 is used to drive the winding wheel 1, one end of the hip joint pulley module 5 is connected to the waist component 4, the other end of the hip joint pulley module 5 is connected to one end of the thigh rod 6, the other end of the thigh rod 6 is connected to one end of the knee joint pulley module 7, and the other end of the knee joint pulley module 7 is connected to one end of the calf rod 8.

[0167] A lower limb exoskeleton sensitivity amplification control device based on adaptive parameter estimation, comprising:

[0168] The first module is used to establish a two-degree-of-freedom dynamic model of a single leg of a lower limb exoskeleton robot;

[0169] The second module is used to perform online estimation of unknown parameters of the established two-degree-of-freedom dynamics model of a single leg of the lower limb exoskeleton robot, and obtain an estimated two-degree-of-freedom dynamics model of the single leg of the lower limb exoskeleton robot;

[0170] The third module is used to substitute the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot into the sensitivity amplification controller to obtain the required control law.

[0171] A processor is used to run a program, wherein when the program is run, any one of the above-mentioned lower limb exoskeleton sensitivity amplification control methods based on adaptive parameter estimation is executed.

[0172] Simulation test:

[0173] Simulate the designed controller in MATLAB and set the parameters of the lower limb exoskeleton: thigh mass m h =0.35kg, calf mass m k =0.30kg, thigh bar length l c,h =0.25m, calf rod length l c,k=0.25m, gravitational acceleration g = 9.8m / s 2 Since the system has only two active degrees of freedom, including the hip joint and the knee joint, discrete points are obtained according to the CGA clinical gait database, and the obtained discrete points are fitted to the equation of time 2s:

[0174]

[0175] Where: q hmh ,q hmk is the trajectory of the hip and knee joints in a gait cycle of an actual human body, that is, Figure 1 Middle hm .

[0176] The designed controller is substituted into the two-degree-of-freedom dynamics model of the single leg of the lower limb exoskeleton robot through the Matlab / Simulink simulation platform to obtain the following Figure 4 , Figure 5 tracking control effect. Figure 4 , Figure 5 It shows that the joint trajectory of the lower limb exoskeleton robot system is basically consistent with the walking trajectory of the human body, achieving a high degree of following (among which, Figure 4 In the figure, Reference HipSignal represents the trajectory of the hip joint of the actual human body in a gait cycle, and Actual Hip Signal represents the actual output trajectory of the hip joint of the lower limb exoskeleton robot system; Figure 5 In the figure, Reference Knee Signal represents the trajectory of the knee joint of an actual human body in a gait cycle, and Actual Knee Signal represents the actual output trajectory of the knee joint of the lower limb exoskeleton robot system). Figure 6 It is shown that the parameter estimation algorithm adopted effectively converges to the true value, which proves the effectiveness of the present invention.

[0177] The specific implementation modes of the present invention are described in detail above in conjunction with the accompanying drawings, but the present invention is not limited to the above implementation modes, and various changes can be made within the knowledge scope of ordinary technicians in this field without departing from the purpose of the present invention.

Claims

1. A method for sensitivity amplification control of lower limb exoskeleton based on adaptive parameter estimation, characterized in that: include: Establish a two-degree-of-freedom dynamic model of a single leg of a lower limb exoskeleton robot; The unknown parameters of the established two-degree-of-freedom dynamic model of the single leg of the lower limb exoskeleton robot are estimated online to obtain the estimated two-degree-of-freedom dynamic model of the single leg of the lower limb exoskeleton robot; Substitute the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot into the sensitivity amplification controller to obtain the required control law; The two-degree-of-freedom dynamic model of a single leg of a lower limb exoskeleton robot is established, comprising: The Lagrangian L is defined as: (1) Where: The kinetic energy of the lower limb exoskeleton robot system, The potential energy of the lower limb exoskeleton robot system; Lagrangian L After taking the partial derivative, we can differentiate the Lagrange equation with respect to time t: (2) Where: are all generalized external force vectors in motion, is the generalized coordinate vector of the system; According to the geometric model of the lower limb exoskeleton, the distance from the center of mass of the thigh rod (12) to the center of rotation of the hip joint (10) and the distance from the center of mass of the calf rod (13) to the center of rotation of the knee joint (11) are obtained as follows: (3) Where: is the distance from the center of mass of the thigh member (12) to the center of rotation of the hip joint (10), is the distance from the mass center of the shank member (13) to the rotation center of the knee joint (11), is the spatial coordinate of the mass center of the thigh member (12), is the spatial coordinate of the mass center of the shank member (13); The velocity of the bar's center of mass is: (4) Where: is the velocity of the center of mass of the thigh member (12), is the velocity of the shank member center of mass (13), is the angular velocity of the hip joint rotation center (10), is the angular velocity of the knee joint rotation center (11); The lower limb exoskeleton robot has two active degrees of freedom per leg, and the kinetic energy of the lower limb exoskeleton robot system is: (5) Where: are the kinetic energy of the thigh rod (6) and the calf rod (8) respectively; , are the masses of the hip joint rotation centers (10) of the thigh rod (6) and the calf rod (8) respectively; , are the moments of inertia of the thigh rod (6) and the calf rod (8) respectively; Angular acceleration of the hip joint rotation center (10) and the knee joint rotation center (11) respectively; are the angle, angular velocity, and angular acceleration of the hip joint rotation center (10), respectively; are the angle, angular velocity, and angular acceleration of the knee joint rotation center (11), respectively; The potential energy of the lower limb exoskeleton robot system is: (6) Where: g is the acceleration due to gravity; Based on the Lagrange equation, the two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot is obtained: (7) Where: is the inertia matrix, is the centripetal force matrix, is the gravity vector; generating actuating torque for actuators of lower limb exoskeleton robotic systems; The method of online estimating the unknown parameters of the established two-degree-of-freedom dynamics model of the single leg of the lower limb exoskeleton robot to obtain the estimated two-degree-of-freedom dynamics model of the single leg of the lower limb exoskeleton robot comprises: performing linear parameterization processing on the obtained two-degree-of-freedom dynamics model of the single leg of the lower limb exoskeleton robot, introducing a first-order low-pass filtering operation, and designing an adaptive law driven by parameter estimation errors to effectively estimate the unknown parameters; The method of online estimating unknown parameters of the established two-degree-of-freedom dynamics model of a single leg of the lower limb exoskeleton robot to obtain an estimated two-degree-of-freedom dynamics model of a single leg of the lower limb exoskeleton robot comprises: Rewrite the left side of the two-degree-of-freedom dynamic model equation of the single leg of the lower limb exoskeleton robot into a linear parameterized form: (8) Where: is the inertia matrix, is the centripetal force matrix, is the gravity vector, inertia matrix, centripetal force matrix, and the gravity vector is the matrix containing unknown parameters in the two-degree-of-freedom dynamics model of the single leg of the lower limb exoskeleton robot; is a , The constant vector to be estimated, is the known regression matrix; Then define the intermediate parameters , ; get: (9) Where: Does not include acceleration terms The regression matrix of The two-degree-of-freedom dynamics model of a single leg of the lower limb exoskeleton robot is rewritten to obtain a reconstructed two-degree-of-freedom dynamics model of a single leg of the lower limb exoskeleton robot: (10) Where: ; Low-pass filtering is introduced on both sides of the two-degree-of-freedom dynamic model equation of the reconstructed lower limb exoskeleton robot single leg to obtain: (11) Where: , They are , The filtered regression matrix; After low-pass filtering ; (12) Where: k represents the low-pass filter coefficient, t represents time; Substituting formula (12) into formula (11), we obtain: (13) Where: is a new regression matrix that does not contain acceleration ; Then define the matrix and vector : (14) Where: It is an adjustable parameter; express The transpose of Respectively represent matrices ,vector The value at the initial time t=0; Defining vectors for: (15) Where: express and The error of Design the corresponding system estimation parameters The adaptive law is: (16) Where: is the learning gain matrix; By estimating Substituting the value back into formula (7), we can obtain the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot: (17) Where: They respectively represent the inertia matrix, centripetal force matrix, and gravity vector of the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot.

2. The method for lower limb exoskeleton sensitivity amplification control based on adaptive parameter estimation according to claim 1, characterized in that: The inertia matrix, centripetal force matrix, gravity vector, and driving torque of the lower limb exoskeleton system are as follows: Where: represents the elements in the inertia matrix; represents the elements in the centripetal force matrix, Represents the elements in the gravity vector.

3. The lower limb exoskeleton sensitivity amplification control method based on adaptive parameter estimation according to claim 1, characterized in that: Substituting the estimated two-degree-of-freedom dynamics model of a single leg of the lower limb exoskeleton robot into the sensitivity amplification controller to obtain the required control law includes: definition It is a model of human-computer interaction. is the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot, is the human-machine interaction force, is the driving torque generated by the actuator of the lower limb exoskeleton robot system, Represents the output trajectory parameters of the lower limb exoskeleton robot system, then: (18) (19) Where: It represents the torque generated by the joints in the lower limb exoskeleton robot system under the human-robot interaction model; The designed sensitivity amplification control law is: (20) Where: is the sensitivity amplification factor; They respectively represent the inertia matrix, centripetal force matrix, and gravity vector of the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot.

4. The lower limb exoskeleton sensitivity amplification control method based on adaptive parameter estimation according to claim 1, characterized in that: The lower limb exoskeleton geometric model is simplified by a lower limb exoskeleton robot system: the hip joint pulley module (5) is simplified to a hip joint rotation center (10), the knee joint pulley module (7) is simplified to a knee joint rotation center (11), the thigh rod mass center (12) is used as the thigh rod mass center (6), and the calf rod mass center (13) is used as the calf rod mass center (8).

5. A lower limb exoskeleton sensitivity amplification control device based on adaptive parameter estimation, used in the lower limb exoskeleton sensitivity amplification control method based on adaptive parameter estimation as claimed in claim 1, characterized in that: include: The first module is used to establish a two-degree-of-freedom dynamic model of a single leg of a lower limb exoskeleton robot; The second module is used to perform online estimation of unknown parameters of the established two-degree-of-freedom dynamics model of a single leg of the lower limb exoskeleton robot, and obtain an estimated two-degree-of-freedom dynamics model of the single leg of the lower limb exoskeleton robot; The third module is used to substitute the estimated two-degree-of-freedom dynamic model of a single leg of the lower limb exoskeleton robot into the sensitivity amplification controller to obtain the required control law.

6. A processor, characterized in that: The processor is used to run a program, wherein the program, when running, executes the lower limb exoskeleton sensitivity amplification control method based on adaptive parameter estimation as described in any one of claims 1 to 4.

Images