CN112388616B - Adaptive robust force control method and device for under-actuated support leg assistance exoskeleton - Google Patents

Adaptive robust force control method and device for under-actuated support leg assistance exoskeleton Download PDF

Info

Publication number
CN112388616B
CN112388616B CN202011363672.0A CN202011363672A CN112388616B CN 112388616 B CN112388616 B CN 112388616B CN 202011363672 A CN202011363672 A CN 202011363672A CN 112388616 B CN112388616 B CN 112388616B
Authority
CN
China
Prior art keywords
exoskeleton
under
support leg
model
actuated support
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202011363672.0A
Other languages
Chinese (zh)
Other versions
CN112388616A (en
Inventor
陈珊
韩腾辉
鹿牧野
王子辛
董方方
韩江
夏链
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hefei University of Technology
Original Assignee
Hefei University of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hefei University of Technology filed Critical Hefei University of Technology
Priority to CN202011363672.0A priority Critical patent/CN112388616B/en
Publication of CN112388616A publication Critical patent/CN112388616A/en
Application granted granted Critical
Publication of CN112388616B publication Critical patent/CN112388616B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00—Programme-controlled manipulators
    • B25J9/0006—Exoskeletons, i.e. resembling a human figure
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00—Programme-controlled manipulators
    • B25J9/16—Programme controls
    • B25J9/1602—Programme controls characterised by the control system, structure, architecture
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00—Programme-controlled manipulators
    • B25J9/16—Programme controls
    • B25J9/1656—Programme controls characterised by programming, planning systems for manipulators
    • B25J9/1664—Programme controls characterised by programming, planning systems for manipulators characterised by motion, path, trajectory planning

Landscapes

  • Engineering & Computer Science (AREA)
  • Robotics (AREA)
  • Mechanical Engineering (AREA)
  • Automation & Control Theory (AREA)
  • Manipulator (AREA)

Abstract

The invention discloses a self-adaptive robust force control method and device for an under-actuated support leg power-assisted exoskeleton. The control method comprises the following steps: establishing a physical model and converting the physical model into a state equation; the physical model comprises a human-computer interface model, a motion model and a complete constraint model; obtaining reference displacement of the under-actuated support leg assistance exoskeleton; obtaining an actual angle value of the exoskeleton and an actual displacement at the back contact; the actual displacement and the reference displacement are used as input quantities, and the output of the lower layer position tracking controller is the motor driving torque; the driving torque of the motor is converted into the output torque of the control current control motor, and the joints of the under-actuated support leg assisting exoskeleton are driven to rotate. Under the condition that the number of the motors is less than the number of the degrees of freedom of the exoskeleton, the invention effectively overcomes the influence of strong multi-joint coupling and model uncertainty of the power-assisted exoskeleton with under-actuated support legs, realizes good follow and power-assisted effects of the power-assisted exoskeleton on human motion, and has strong application value.

Description

Adaptive robust force control method and device for under-actuated support leg assistance exoskeleton
Technical Field
The invention relates to a control method of an exoskeleton, belongs to the technical field of wearable technology, and particularly relates to a self-adaptive robust force control method of an under-actuated support leg power-assisted exoskeleton and a self-adaptive robust force control device of the under-actuated support leg power-assisted exoskeleton.
Background
The wearable lower limb assistance exoskeleton robot is an intelligent man-machine integrated device which simulates the structure of a human lower limb and enhances the walking durability, walking speed, load bearing capacity and other performances of a wearer, and plays an important role in rescue and relief work, building operation, improvement of individual combat capacity and the like. The combination of exoskeleton and human can adapt to unstructured environments, has excellent flexibility, and can complete some complex tasks, which cannot be compared with other complete mechanical devices.
The fully-driven exoskeleton system has the problems of over-weight and over-energy consumption due to the inclusion of a plurality of drivers, which can limit the load bearing capacity of the system and the cruising capacity of the portable energy supply system. In order to further reduce the weight and energy consumption of the power-assisted exoskeleton robot and enhance the flexibility of human body movement, an under-driven power-assisted exoskeleton robot is gradually proposed. Compared with the fully-actuated exoskeleton, the control method of the fully-actuated exoskeleton cannot be directly used in the system of the fully-actuated exoskeleton due to the lack of control input. In addition, the multi-joint under-actuated exoskeleton system has strong coupling high-order nonlinearity and various model uncertainties, so that the robust performance requirement on the control algorithm is high. The existing underactuated exoskeleton control methods only relate to the primary control of the underactuated exoskeleton, and the dynamic model or the control algorithm is simplified, so that the robustness of the system is not strong.
Disclosure of Invention
In order to solve the technical problem that the system robustness performance of the existing control method of the under-actuated exoskeleton is not strong, the invention provides a self-adaptive robust force control method and a device of the under-actuated support leg power-assisted exoskeleton.
The invention is realized by adopting the following technical scheme: an adaptive robust force control method for an under-actuated support leg assisted exoskeleton, the under-actuated support leg assisted exoskeleton comprising:
a foot portion;
the bottom end of the first rod piece is connected with the foot part;
the bottom end of the second rod piece is rotatably connected with the top end of the first rod piece;
the knee joint motor is used for driving the first connecting rod and the second rod piece to rotate relatively;
a knee joint encoder mounted on the knee joint motor;
the back part is rotatably connected with the top end of the second rod piece;
a hip joint motor for driving the back and the second rod to rotate relatively;
a hip joint encoder mounted on the hip joint motor;
a force sensor mounted on the back;
a back strap, both ends of which are connected with the upper end of the back;
a waist band, both ends of which are connected with the lower end of the back part; and
a real-time controller electrically connected to the knee joint motor, the hip joint motor, the knee joint encoder, the hip joint encoder, and the force sensor;
the adaptive robust force control method comprises the following steps:
(1) initializing a sampling period of the real-time controller;
(2) rotating the foot to horizontal and the first bar, the second bar and the back to vertical positions, and initializing the knee joint encoder and the hip joint encoder and zeroing the encoder values;
(3) initializing the force sensor and zeroing the value of the force sensor;
(4) establishing a physical model of the under-actuated support leg assisting exoskeleton and converting the physical model into a state equation; wherein the physical model comprises a human-computer interface model, a motion model of the exoskeleton mechanical body and a complete constraint model provided by a wearer;
(5) the force sensor is connected with a wearer through the back bandage, acting force on the force sensor is measured, and reference displacement of the under-actuated support leg assisting exoskeleton is obtained through the upper layer controller;
(6) obtaining an actual angle value of the under-actuated support leg assistance exoskeleton through the knee joint encoder and the hip joint encoder, and obtaining an actual displacement of the back contact part according to a positive kinematics model of an exoskeleton system; according to the reference displacement obtained in the step (5), taking the actual displacement and the reference displacement as input quantities of a lower-layer position tracking controller, wherein the output of the lower-layer position tracking controller is motor driving torque of knee joints and hip joints in the under-actuated support leg assisting exoskeleton;
(7) converting the motor driving torque obtained in the step (6) into control current of a corresponding motor through the drivers of the knee joint motor and the hip joint motor; and
(8) and controlling the corresponding output torque of the knee joint motor and the hip joint motor through each control current to drive each joint of the under-actuated support leg assistance exoskeleton to rotate, so as to realize the following motion of the under-actuated support leg assistance exoskeleton.
The ankle joint in the exoskeleton system is driven passively, and a driver is not arranged, so that the exoskeleton system has lighter weight, better cruising ability of a portable energy supply system and higher load performance. The sensor system mainly realizes effective and reliable man-machine interaction by a force sensor and a rotary encoder, and the three-degree-of-freedom underactuated exoskeleton system is converted into a two-degree-of-freedom full-actuated system by considering complete constraint provided by a wearer aiming at the problems of force boosting and following of the underactuated supporting leg assisting exoskeleton. The control method adopts a force control method, utilizes a multi-input multi-output adaptive robust control Algorithm (ARC) to design an upper layer controller and a lower layer controller, effectively overcomes the influence of multi-joint strong coupling and model uncertainty of the power-assisted exoskeleton with under-actuated support legs, solves the technical problem that the system robustness performance of the existing exoskeleton control method is not strong, realizes good following and power-assisted effects of the power-assisted exoskeleton on human motion, and has strong application value.
As a further improvement of the above scheme, the human-machine interface model is:
Figure GDA0003392051300000031
wherein, Fhm＝[Fhmx Fhmy τez]TFor human acting force, T is the sampling period, x, y and z are three-dimensional coordinate axis symbols, tauezIs the moment; k ═ diag { K ═ Kx,Ky,KzRigidity of the human-machine interface, xh＝[xhx xhy xhz]TAnd xe＝[xex xey xez]TDisplacement of the wearer at the back contact and displacement of the exoskeleton, respectively;
Figure GDA0003392051300000041
model uncertainty and interference are concentrated on the human-machine interface;
by integration of the human-machine forces in the transformation of the physical model
Figure GDA0003392051300000042
In place of FhmObtaining the state equation as follows:
Figure GDA0003392051300000043
as a further improvement of the above scheme, the motion model is:
Figure GDA0003392051300000044
in the formula, Fhm＝[Fhmx Fhmy τez]TFor man-machine forces of contact points, τactFor the motor drive torque, J is the Jacobian matrix of the system at the force sensor, q (t) [ q [ q ]) ]1(t),q2(t),q3(t)]TThe rotation angles of the ankle joint, knee joint and hip joint, Msp3(q) is the inertial matrix of the system,
Figure GDA00033920513000000410
is the centrifugal and Coriolis forces of the systemMatrix, Gsp3(q) is the gravity matrix of the system, B is the damping matrix of the system,
Figure GDA0003392051300000045
is the centralized modeling error of the system;
the motion model can be further converted into:
Figure GDA0003392051300000046
in the formula (I), the compound is shown in the specification,
Figure GDA0003392051300000047
as a further improvement of the above scheme, the complete constraint model is:
xez＝xezd(t)
taking the second derivative of the complete constraint model:
Figure GDA0003392051300000048
find out
Figure GDA0003392051300000049
And τezFinally, the following can be obtained:
Figure GDA0003392051300000051
Figure GDA0003392051300000052
in the formula, xea＝[xex xey]T,Mea＝u1Mxu4,Bea＝u1J-Tu3,Cea＝u1Cxu4,Gea＝u1Gx,Bxea＝u1Bxu4,Mea2＝u2Mxu4,Cea2＝u2Cxu4,Gea2＝u2Gx,Bxea2＝u2Bxu4,Bea2＝-u2J-Tu3,
Figure GDA0003392051300000053
u2＝[0 0 1],
Figure GDA0003392051300000054
As a further improvement of the above solution, the method for converting the physical model into the equation of state comprises the following steps:
(4.1) order State variables
Figure GDA0003392051300000055
Wherein Fh'm＝[Fhmx Fhmy]T,
Figure GDA0003392051300000056
x2a＝xea,
Figure GDA0003392051300000057
Let the centralized model uncertainty be:
Figure GDA0003392051300000058
(4.2) dividing the uncertainty of the centralized model into a constant part and a time-varying function part to obtain
Figure GDA0003392051300000059
i＝1a,3a，ΔinAnd ΔiRespectively represent
Figure GDA00033920513000000510
Constant part and time-varying part of；
(4.3) is provided with
Figure GDA00033920513000000511
Wherein, Kθa＝[1/Kx1/Ky]T，Δ1an＝[Δ1anx Δ1any]T，β＝[Y2 Y3 Y4 X4 J2 J3 J4]TSystem parameters for exoskeleton support legs, Bθ＝[B1 B2 B3]T，Δ3an＝[Δ3anx Δ3any]T。
As a further improvement of the above scheme, the state equation of the physical model of the under-actuated support leg assisted exoskeleton is as follows:
Figure GDA00033920513000000512
Figure GDA00033920513000000513
Figure GDA0003392051300000061
wherein, Kxy＝diag{Kx,Ky}。
As a further improvement of the above aspect, the control method of the upper controller includes the steps of:
according to the state equation of the physical model in the step (4), setting the first tracking error as z1a＝x1a-x1adWherein x is1adThe integral of the expected man-machine acting force in the x and y directions is 0; let xmFor the first virtual control input, the first virtual control input xmFirst tracking error z for man-machine effort1aRapidly towards zero;
let xm＝xma+xms+xmsnWherein
Figure GDA0003392051300000062
xms＝K1z1a，fθFAnd YθFIs formed by xmaIs linearized with respect to the parameters of (f)θF＝[0 0]T，
Figure GDA0003392051300000063
Desired human-machine forces in the x, y directions, respectively, K1＝diag{K1x,K1yIs a linear feedback gain matrix and,
Figure GDA0003392051300000064
is thetaFAnd the range of the estimated values is:
Figure GDA0003392051300000065
wherein
Figure GDA0003392051300000066
To a parameter thetaFIs estimated value of
Figure GDA0003392051300000067
The minimum value of (a) is determined,
Figure GDA0003392051300000068
to a parameter thetaFIs estimated value of
Figure GDA0003392051300000069
Maximum value of (d); estimated value
Figure GDA00033920513000000610
In the upper layer controller by adaptive rate
Figure GDA00033920513000000611
Is obtained in which
Figure GDA00033920513000000612
Γ1Is a matrix of positive fixed gains, and,
Figure GDA00033920513000000613
the mapping function of (d) is:
Figure GDA00033920513000000614
in the formula-iIs an independent variable; x is the number ofmsnSatisfies the following conditions:
Figure GDA00033920513000000615
Figure GDA00033920513000000616
in the formula (I), the compound is shown in the specification,
Figure GDA00033920513000000617
is an estimated value
Figure GDA00033920513000000618
Minus the actual value thetaF，
Figure GDA00033920513000000619
ε1Is a threshold and is arbitrarily non-negative;
according to a first virtual control input xmiThe reference displacement, the reference speed and the reference acceleration of the exoskeleton are obtained by smoothing the reference displacement, the reference speed and the reference acceleration through a third-order filter, wherein i is 1 and 2; wherein, the state equation of the third-order filter is:
Figure GDA0003392051300000071
Figure GDA0003392051300000072
Figure GDA0003392051300000073
i＝1,2
let yiRepresents the exoskeleton reference displacement, let yi(s)＝xmi(s)，xi(1),xi(2),xi(3) Respectively represent the filtered reference displacement, reference velocity and reference acceleration, then yiTo xi(1) The transfer function is:
Figure GDA0003392051300000074
obtaining x by the transfer functionmiConverting into required smooth exoskeleton reference displacement xi(1) (ii) a Wherein, a1,a2,a3Obtained by pole placement.
As a further improvement of the above solution, the method for designing the lower layer position tracking controller includes the steps of:
setting a second tracking error
Figure GDA0003392051300000075
Wherein
Figure GDA0003392051300000076
Defining a conversion equation:
Figure GDA0003392051300000077
wherein, K2Taking an arbitrary non-negative number, z3Is the third tracking error; z is a radical of2And z3Has a transfer function of
Figure GDA0003392051300000078
Let Bxeax3a＝YBBθ，
Figure GDA0003392051300000079
Wherein β ═ Y2 Y3 Y4 X4 J2 J3 J4]T,Y2,Y3,Y4,X4,J2,J3,J4Model parameters, B, both of mechanical structureθ＝[B1 B2 B3]TDamping of the system; control of tauactThe determination method comprises the following steps: tau isact＝τacta+τacts+τactsnWherein
Figure GDA00033920513000000710
K3Is a linear feedback gain that is a function of,
Figure GDA00033920513000000711
are respectively beta, Bθ，Δ3anIs determined by the estimated value of (c),
Figure GDA00033920513000000712
wherein
Figure GDA00033920513000000713
To a parameter thetaqIs estimated value of
Figure GDA00033920513000000714
The minimum value of (a) is determined,
Figure GDA0003392051300000081
to a parameter thetaqIs estimated value of
Figure GDA0003392051300000082
Maximum value of (d); estimated value
Figure GDA0003392051300000083
Is controlled by an adaptation rate in the lower layer position tracking controller
Figure GDA0003392051300000084
Obtained byIn
Figure GDA0003392051300000085
Γ2Is a matrix of positive fixed gains, and,
Figure GDA0003392051300000086
the mapping function of (d) is:
Figure GDA0003392051300000087
in the formula-iIs an independent variable;
order to
Figure GDA0003392051300000088
τactsnSatisfies the following conditions:
Figure GDA0003392051300000089
Figure GDA00033920513000000810
in the formula (I), the compound is shown in the specification,
Figure GDA00033920513000000811
is an estimated value
Figure GDA00033920513000000812
Minus the actual value thetaq，
Figure GDA00033920513000000813
ε3Is a threshold and is arbitrarily non-negative.
As a further improvement of the above scheme, the back part is a back plate, and the knee joint encoder and the hip joint encoder are joint rotary encoders; the under-actuated support leg assistance exoskeleton further comprises:
the knee joint motor drives the first connecting rod and the second rod piece to rotate relatively through the knee joint reducer;
and the hip joint motor drives the back and the second rod piece to rotate relatively through the hip joint reducer.
The invention also provides a self-adaptive robust force control device of the under-actuated support leg assisting exoskeleton, which applies any one of the self-adaptive robust force control methods of the under-actuated support leg assisting exoskeleton, and comprises the following steps:
the initialization module I is used for initializing a sampling period of the real-time controller, and the sampling period is between 10 and 20 milliseconds;
a second initialization module, configured to rotate the foot to a horizontal position, rotate the first rod, the second rod, and the back to a vertical position, initialize the knee joint encoder and the hip joint encoder, and zero an encoder value;
the initialization module III is used for initializing the force sensor and zeroing the numerical value of the force sensor;
the model establishing module is used for establishing a physical model of the under-actuated support leg assistance exoskeleton and converting the physical model into a state equation; wherein the physical model comprises a human-computer interface model, a motion model of the exoskeleton mechanical body and a complete constraint model provided by a wearer;
the reference displacement acquisition module is used for connecting the force sensor with a wearer through the back bandage, measuring acting force on the force sensor and acquiring reference displacement of the under-actuated support leg assistance exoskeleton through the upper layer controller;
the actual displacement acquisition module is used for acquiring an actual angle value of the under-actuated support leg assistance exoskeleton through the knee joint encoder and the hip joint encoder and acquiring an actual displacement of the back contact part according to a positive kinematics model of an exoskeleton system; the actual displacement acquisition module takes the actual displacement and the reference displacement as input quantities of a lower-layer position tracking controller according to the reference displacement, and the output of the lower-layer position tracking controller is motor driving torque at knee joints and hip joints in the under-actuated support leg assisting exoskeleton;
the conversion module is used for converting the driving torque of the motors into control currents of the corresponding motors through the drivers of the knee joint motor and the hip joint motor; and
and the following module is used for controlling the corresponding output torque of the knee joint motor and the hip joint motor through each control current to drive each joint of the under-actuated support leg assistance exoskeleton to rotate so as to realize the following motion of the under-actuated support leg assistance exoskeleton.
Compared with the existing exoskeleton control method, the adaptive robust force control method and device for the under-actuated support leg assisting exoskeleton have the following beneficial effects:
1. according to the self-adaptive robust force control method of the under-actuated support leg assistance exoskeleton, ankle joints in an exoskeleton system are driven passively, and a driver is not arranged, so that the exoskeleton has lighter weight, better cruising ability of a portable energy supply system and higher load performance.
2. According to the adaptive robust force control method for the under-actuated support leg power-assisted exoskeleton, a sensor system of the method is mainly characterized in that a force sensor and a rotary encoder are used for realizing effective and reliable man-machine interaction, and the complete constraint provided by a wearer is considered for solving the problems of force reinforcement and following of the under-actuated support leg power-assisted exoskeleton, so that the three-degree-of-freedom under-actuated exoskeleton system is converted into a two-degree-of-freedom full-actuation system.
3. The adaptive robust force control method for the under-actuated support leg power-assisted exoskeleton adopts a force control method, utilizes a multi-input multi-output adaptive robust control Algorithm (ARC) to design an upper layer controller and a lower layer controller, effectively overcomes the influences of strong multi-joint coupling and model uncertainty of the under-actuated support leg power-assisted exoskeleton under the condition that the number of motors is less than the number of exoskeleton motion freedom degrees, performs feedforward compensation on a control model to ensure zero tracking error under a static state, ensures the dynamic characteristic and stability of the under-actuated power-assisted exoskeleton system through designed robust feedback, solves the technical problem that the system robustness of the existing exoskeleton control method is not strong, realizes good following and power-assisted effects of the power-assisted exoskeleton on human motion, and has strong application value.
4. The self-adaptive robust force control method of the under-actuated support leg assisting exoskeleton fully considers the control effect of a wearer on exoskeletons, reduces the use of a motor and the consumption of energy, is effective and reliable in the problem of man-machine interaction, and has the characteristic of quick response to the movement intention of a human body.
5. According to the self-adaptive robust force control method for the under-actuated support leg assisting exoskeleton, an exoskeleton wearer is used as a participant of system control, the wearer can guarantee the front-back walking balance of the whole system in a walking plane, and the exoskeleton is prevented from falling down. Meanwhile, the method utilizes a cascade control strategy to design upper and lower layer controllers, realizes the trajectory planning and trajectory tracking of the under-actuated power-assisted exoskeleton, and is simple to realize, easy to realize in engineering and flexible to control.
6. The beneficial effects of the adaptive robust force control device of the under-actuated support leg power-assisted exoskeleton are the same as those of the adaptive robust force control method of the under-actuated support leg power-assisted exoskeleton, and the detailed description is omitted here.
Drawings
Fig. 1 is a schematic diagram of the overall shape and structure of an under-actuated support leg assisted exoskeleton applied to an adaptive robust force control method for the under-actuated support leg assisted exoskeleton in embodiment 1 of the present invention.
Fig. 2 is a front view of the under-actuated support leg assisted exoskeleton of fig. 1.
Figure 3 is a side view of the under-actuated support leg assisted exoskeleton of figure 1.
Fig. 4 is a control block diagram of an adaptive robust force control method of an under-actuated support leg assisted exoskeleton in embodiment 1 of the present invention.
Fig. 5 is a control flowchart of an adaptive robust force control method for an under-actuated support leg assisted exoskeleton in embodiment 1 of the invention.
Description of the symbols:
1 foot 8 hip joint motor
2 first rod 9 hip joint reducer
3 Knee joint encoder 10 Back
4 knee joint motor 11 force sensor
12 back bandage of knee joint reduction gear
6 second rod 13 waist strap
7 hip joint encoder
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Example 1
Referring to fig. 1, fig. 2 and fig. 3, the present embodiment provides an adaptive robust force control method for an under-actuated support leg assisted exoskeleton, which is used to control an under-actuated support leg assisted exoskeleton. The under-actuated support leg assisting exoskeleton comprises a foot part 1, a first rod part 2, a second rod part 6, a knee joint motor 4, a knee joint reducer 5, a knee joint encoder 3, a back part 10, a hip joint motor 8, a hip joint reducer 9, a hip joint encoder 7, a force sensor 11, a back bandage 12, a waist bandage 13 and a real-time controller (not shown in the figure). The knee joint encoder 3 and the hip joint reducer 9 are joint rotary encoders, and the back 10 is a back plate.
The foot 1 is used as a contact part of the under-actuated support leg assisted exoskeleton and the ground, and can support the whole exoskeleton to function similarly to the feet of a human body. The bottom end of the first rod 2 is connected with the foot 1, and can be movably connected or connected in other connection modes. The bottom end of the second rod 6 is rotatably connected (can be connected by a hinge) with the top end of the first rod 2, and the knee joint motor 4 is used for driving the first connecting rod and the second rod 6 to rotate relatively. The knee joint encoder 3 is installed on the knee joint motor 4 (can be disposed at the position connected by the hinge), and in this embodiment, the knee joint motor 4 drives the first link and the second link 6 to rotate relatively through the knee joint reducer 5. The back 10 is connected with the top end of the second rod 6 in a rotating way (can be connected through a hinge), and the hip joint motor 8 is used for driving the back 10 and the second rod 6 to rotate relatively. The hip encoder 7 is mounted on a hip motor 8 (which may be arranged at the hinged connection), and in this embodiment the hip motor 8 drives the back 10 and the second bar 6 in rotation relative to each other via a hip reducer 9. The force sensor 11 is mounted on the back 10, both ends of the back strap 12 are connected to the upper end of the back 10, and both ends of the waist strap 13 are connected to the lower end of the back 10. The real-time controller is electrically connected with the knee joint motor 4, the hip joint motor 8, the knee joint encoder 3, the hip joint encoder 7 and the force sensor 11. The real-time controller may be a product of type NI cRIO-9031, but is not limited thereto.
Referring to fig. 4 and 5, based on the under-actuated support leg assisting exoskeleton, in order to overcome the second-order non-integrity of the under-actuated assisting exoskeleton caused by the fact that the number of drivers is less than the number of system degrees of freedom, in the embodiment, an exoskeleton wearer is used as a participant of system control, and in a walking plane, the wearer can ensure the front-back walking balance of the whole system, so that the exoskeleton is prevented from falling down. Thus, it is assumed that the wearer can provide a counterbalancing moment that rotates about the z-axis so that the rotational angle of the exoskeleton back plate follows a bounded trajectory. Considering the complete constraint provided by the wearer, the three-degree-of-freedom underactuated system of a joint angle space is finally changed into a two-degree-of-freedom full-actuated system related to the Cartesian position of the exoskeleton back plate. In order to overcome the influence of uncertainty in the modeling process of the under-actuated assisting exoskeleton and achieve good follow-up and assisting effects of the assisting exoskeleton on human motion, the control strategy of the under-actuated supporting leg assisting exoskeleton in the embodiment adopts Adaptive Robust Control (ARC) which can well overcome the influence of model uncertainty. The principle of Adaptive Robust Control (ARC) is to design adaptive rate to continuously adjust model parameters, to perform feedforward compensation on a control model to ensure zero tracking error under static state, and to ensure dynamic characteristics and stability of the under-actuated power-assisted exoskeleton system through designed robust feedback. Meanwhile, the upper layer controller and the lower layer controller are designed by utilizing a cascade control strategy, the track planning and the track tracking of the under-actuated power-assisted exoskeleton are realized, the control algorithm is simple to realize, the engineering is easy to realize, and the control is flexible. Therefore, in particular, the adaptive robust force control method in the present embodiment includes the following steps.
(1) The sampling period of the real-time controller is initialized. In the present embodiment, the value of the sampling period T is between 10 and 20 milliseconds.
(2) The foot 1 is rotated to the horizontal and the first bar 2, the second bar 6 and the back 10 are rotated to the vertical position and the knee joint encoder 3 and the hip joint encoder 7 are initialized and the encoder values are zeroed.
(3) The force sensor 11 is initialized and the value of the force sensor 11 is zeroed.
(4) And establishing a physical model of the under-actuated support leg assistance exoskeleton and converting the physical model into a state equation. The physical model comprises a human-computer interface model, a motion model of the exoskeleton mechanical body and a complete constraint model provided by a wearer.
The human-computer interface model is as follows:
Figure GDA0003392051300000131
wherein, Fhm＝[Fhmx Fhmy τez]TFor human acting force, T is sampling period, x, y and z are three-dimensional coordinate axis symbols, tauezIs a moment. K ═ diag { K ═ Kx,Ky,KzRigidity of the human-machine interface, xh＝[xhx xhy xhz]TAnd xe＝[xexxey xez]TRespectively displacement of the wearer at the contact of the back 10 and displacement of the exoskeleton.
Figure GDA0003392051300000132
Is a centralized model uncertainty and disturbance on the human-machine interface.
In transforming the physical model, the human-machine interface model is a static equation, so Fhm、xhAnd xeThe relationship between them is static, in order to allow dynamic control of the man-machine forces FhmBy integration of man-machine forces
Figure GDA0003392051300000133
In place of FhmThe equation of state is obtained as:
Figure GDA0003392051300000134
the motion model is as follows:
Figure GDA0003392051300000141
in the formula, Fhm＝[Fhmx Fhmy τez]TFor man-machine forces of contact points, τactFor the motor drive torque, J is the jacobian matrix of the system at the force sensor 11, q (t) ═ q1(t),q2(t),q3(t)]TThe rotation angles of the ankle joint, knee joint and hip joint, Msp3(q) is the inertial matrix of the system,
Figure GDA0003392051300000142
is the centrifugal and Coriolis force matrix of the system, Gsp3(q) is the gravity matrix of the system, B is the damping matrix of the system,
Figure GDA0003392051300000143
is the centralized modeling error of the system.
Because:
Figure GDA0003392051300000144
Figure GDA0003392051300000145
the motion model can be further converted into:
Figure GDA0003392051300000146
in the formula (I), the compound is shown in the specification,
Figure GDA0003392051300000147
due to the lack of control input to the under-actuated, assisted exoskeleton system, the wearer needs to provide a certain control torque to ensure the stability of the whole system. In the walking plane, the wearer can ensure the front and back walking balance of the whole system, and the exoskeleton is prevented from falling down. Thus, it is assumed that the wearer is able to provide a counterbalancing moment τ about the z-axisezSo that the rotation angle of the exoskeleton back plate follows a bounded track xezd(t) motion, i.e. the complete constraint model provided by the wearer:
xez＝xezd(t)
the second derivative is obtained by calculating
Figure GDA0003392051300000148
By combining the motion models, the complete constraint model can know that four unknowns exist
Figure GDA0003392051300000151
Four correlation equations, so in τactFor input, can find
Figure GDA0003392051300000152
And τezFinally, the following can be obtained:
Figure GDA0003392051300000153
Figure GDA0003392051300000154
in the formula, xea＝[xex xey]T,Mea＝u1Mxu4,Bea＝u1J-Tu3,Cea＝u1Cxu4,Gea＝u1Gx,Bxea＝u1Bxu4,Mea2＝u2Mxu4,Cea2＝u2Cxu4,Gea2＝u2Gx,Bxea2＝u2Bxu4,Bea2＝-u2J-Tu3,
Figure GDA0003392051300000155
u2＝[0 0 1],
Figure GDA0003392051300000156
The method for converting the physical model into the state equation comprises the following steps:
(4.1) order State variables
Figure GDA0003392051300000157
Wherein F'hm＝[Fhmx Fhmy]T,
Figure GDA0003392051300000158
x2a＝xea,
Figure GDA0003392051300000159
Let the centralized model uncertainty be:
Figure GDA00033920513000001510
(4.2) dividing the uncertainty of the centralized model into a constant part and a time-varying function part to obtain
Figure GDA00033920513000001511
i＝1a,3a，ΔinAnd ΔiRespectively represent
Figure GDA00033920513000001512
A constant portion and a time-varying portion of;
(4.3) is provided with
Figure GDA00033920513000001513
Wherein, Kθa＝[1/Kx 1/Ky]T，Δ1an＝[Δ1anx Δ1any]T，β＝[Y2 Y3 Y4 X4 J2 J3 J4]TSystem parameters for exoskeleton support legs, Bθ＝[B1 B2 B3]T，Δ3an＝[Δ3anx Δ3any]T。
Thus, the state equation of the physical model of the under-actuated support leg assistance exoskeleton is as follows:
Figure GDA00033920513000001514
Figure GDA00033920513000001515
Figure GDA0003392051300000161
wherein, Kxy＝diag{Kx,Ky}。
(5) The force sensor 11 is connected with a wearer through the back bandage 12, the acting force on the force sensor 11 is measured, and the reference displacement of the under-actuated support leg assisting exoskeleton is obtained through the upper layer controller. The control method of the upper layer controller comprises the following steps:
according to the state equation of the physical model in the step (4), setting the first tracking error as z1a＝x1a-x1adWherein x is1adThe integral of the expected man-machine acting force in the x and y directions is 0; let xmFor the first virtual control input, the first virtual control input xmFirst tracking error z for man-machine effort1aRapidly towards zero;
let xm＝xma+xms+xmsnWherein
Figure GDA0003392051300000162
xms＝K1z1a，fθFAnd YθFIs formed by xmaIs linearized with respect to the parameters of (f)θF＝[0 0]T，
Figure GDA0003392051300000163
Desired human-machine forces in the x, y directions, respectively, K1＝diag{K1x,K1yIs a linear feedback gain matrix. In this example, take K1＝diag{13,20}。
Figure GDA0003392051300000164
Is thetaFAnd the range of the estimated values is:
Figure GDA0003392051300000165
in this embodiment, the initial value is
Figure GDA0003392051300000166
Wherein
Figure GDA0003392051300000167
To a parameter thetaFIs estimated value of
Figure GDA0003392051300000168
Is taken in this embodiment
Figure GDA0003392051300000169
Figure GDA00033920513000001617
To a parameter thetaFIs estimated value of
Figure GDA00033920513000001610
In this embodiment, take
Figure GDA00033920513000001611
Estimated value
Figure GDA00033920513000001612
In the upper controller by adaptive rate
Figure GDA00033920513000001613
Is obtained in which
Figure GDA00033920513000001614
Γ1Is a positive fixed gain matrix, in this embodiment, taking Γ1＝diag{0,0,100,100}，
Figure GDA00033920513000001615
The mapping function of (d) is:
Figure GDA00033920513000001616
in the formula-iIs an independent variable.
According to an Adaptive Robust (ARC) control algorithm, xmsnSatisfies the following conditions:
Figure GDA0003392051300000171
Figure GDA0003392051300000172
in the formula (I), the compound is shown in the specification,
Figure GDA0003392051300000173
is an estimated value
Figure GDA0003392051300000174
Minus the actual value thetaF，
Figure GDA0003392051300000175
ε1Is a threshold and is arbitrarily non-negative. In the present embodiment,. epsilon.1＝[1 1]TSelecting xmsn＝[0 0]T。
According to a first virtual control input xmiThe reference displacement, the reference speed and the reference acceleration of the exoskeleton are obtained by smoothing the reference displacement, the reference speed and the reference acceleration through a third-order filter, wherein i is 1 and 2; wherein, the state equation of the third-order filter is:
Figure GDA0003392051300000176
Figure GDA0003392051300000177
Figure GDA0003392051300000178
i＝1,2
let yiRepresents the exoskeleton reference displacement, let yi(s)＝xmi(s)，xi(1),xi(2),xi(3) Respectively represent the filtered reference displacement, reference velocity and reference acceleration, then yiTo xi(1) The transfer function is:
Figure GDA0003392051300000179
by means of a transfer function, x is obtainedmiConverting into required smooth exoskeleton reference displacement xi(1) (ii) a Wherein, a1,a2,a3Obtained by pole placement. In this embodiment, the closed loop pole is set to 20 radian per second to obtain a1,a2,a3,a4Respectively is a1＝80,a2＝2400,a332000, this may not be limiting.
(6) Actual angle values of the under-actuated support leg assistance exoskeleton are obtained through the knee joint encoder 3 and the hip joint encoder 7, and actual displacement of a contact position of the back 10 is obtained according to a positive kinematics model of the exoskeleton system. And (5) according to the reference displacement obtained in the step (5), taking the actual displacement and the reference displacement as input quantities of a lower-layer position tracking controller, wherein the output of the lower-layer position tracking controller is motor driving torque at knee joints and hip joints in the under-actuated support leg assisting exoskeleton.
In this embodiment, the method for designing the lower layer position tracking controller includes the following steps:
setting a second tracking error
Figure GDA0003392051300000181
Wherein
Figure GDA0003392051300000182
Defining a conversion equation:
Figure GDA0003392051300000183
wherein, K2Taking an arbitrary non-negative number, z3Is the third tracking error. z is a radical of2And z3Has a transfer function of
Figure GDA0003392051300000184
The transfer function is a stable transfer function, soTo make z be2Very small or zero is taken to mean z3Small or near zero. The design of the lower layer position tracking controller is to let z3The dynamic performance is ensured to be as small as possible.
Let Bxeax3a＝YBBθ，
Figure GDA0003392051300000185
Wherein β ═ Y2 Y3 Y4 X4 J2 J3J4]T,Y2,Y3,Y4,X4,J2,J3,J4Model parameters, B, both of mechanical structureθ＝[B1 B2 B3]TIs the damping of the system. Control of tauactThe determination method comprises the following steps: tau isact＝τacta+τacts+τactsnWherein
Figure GDA0003392051300000186
K3Is a linear feedback gain, in this embodiment, K is taken3＝[12000 12000]T。
Figure GDA0003392051300000187
Are respectively beta, Bθ，Δ3anIs determined by the estimated value of (c),
Figure GDA0003392051300000188
in this embodiment, the initial value is taken as
Figure GDA0003392051300000189
Figure GDA00033920513000001810
From the physical model, the range of estimated values that can be obtained is:
Figure GDA00033920513000001811
wherein
Figure GDA00033920513000001812
To a parameter thetaqIs estimated value of
Figure GDA00033920513000001813
The minimum value of (a) is determined,
Figure GDA00033920513000001814
to a parameter thetaqIs estimated value of
Figure GDA00033920513000001815
Is measured. In this embodiment, the selection range is
Figure GDA00033920513000001816
Estimated value
Figure GDA00033920513000001817
Is controlled by the adaptation rate in the underlying position tracking controller
Figure GDA00033920513000001818
Is obtained in which
Figure GDA00033920513000001819
Γ2Is a positive constant gain matrix, in this embodiment, chosen to be Γ2＝diag{100,0,0,0,100,0,0,0,0,0,100,100}，
Figure GDA00033920513000001820
The mapping function of (d) is:
Figure GDA0003392051300000191
in the formula-iIs an independent variable.
Order to
Figure GDA0003392051300000192
According to an Adaptive Robust (ARC) control algorithm, τactsnThe following two conditions are satisfied:
Figure GDA0003392051300000193
Figure GDA0003392051300000194
in the formula (I), the compound is shown in the specification,
Figure GDA0003392051300000195
is an estimated value
Figure GDA0003392051300000196
Minus the actual value thetaq，
Figure GDA0003392051300000197
ε3Is a threshold and is arbitrarily non-negative. In this embodiment, ε is selected3＝[1 1]TSelecting τactsn＝[0 0]T。
(7) The motor driving torque tau obtained in the step (6) is driven by the drivers of the knee joint motor 4 and the hip joint motor 8actAnd converted into the control current of the corresponding motor.
(8) The output torque of the corresponding knee joint motor 4 and the corresponding hip joint motor 8 is controlled through each control current to drive each joint of the under-actuated support leg assisting exoskeleton to rotate, and the following movement of the under-actuated support leg assisting exoskeleton is realized.
In summary, compared with the existing exoskeleton control method, the adaptive robust force control method for the under-actuated support leg assisting exoskeleton of the embodiment has the following advantages:
1. according to the self-adaptive robust force control method of the under-actuated support leg assistance exoskeleton, ankle joints in an exoskeleton system are driven passively, and a driver is not arranged, so that the exoskeleton has lighter weight, better cruising ability of a portable energy supply system and higher load performance.
2. According to the adaptive robust force control method for the under-actuated support leg power-assisted exoskeleton, a sensor system of the method is mainly characterized in that a force sensor 11 and a rotary encoder are used for realizing effective and reliable man-machine interaction, and the three-degree-of-freedom under-actuated exoskeleton system is converted into a two-degree-of-freedom full-drive system by considering complete constraint provided by a wearer aiming at the problems of force boosting and following of the under-actuated support leg power-assisted exoskeleton.
3. The adaptive robust force control method for the under-actuated support leg power-assisted exoskeleton adopts a force control method, utilizes a multi-input multi-output adaptive robust control Algorithm (ARC) to design an upper layer controller and a lower layer controller, effectively overcomes the influences of strong multi-joint coupling and model uncertainty of the under-actuated support leg power-assisted exoskeleton under the condition that the number of motors is less than the number of exoskeleton motion freedom degrees, performs feedforward compensation on a control model to ensure zero tracking error under a static state, ensures the dynamic characteristic and stability of the under-actuated power-assisted exoskeleton system through designed robust feedback, solves the technical problem that the system robustness of the existing exoskeleton control method is not strong, realizes good following and power-assisted effects of the power-assisted exoskeleton on human motion, and has strong application value.
4. The self-adaptive robust force control method of the under-actuated support leg assisting exoskeleton fully considers the control effect of a wearer on exoskeletons, reduces the use of a motor and the consumption of energy, is effective and reliable in the problem of man-machine interaction, and has the characteristic of quick response to the movement intention of a human body.
5. According to the self-adaptive robust force control method for the under-actuated support leg assisting exoskeleton, an exoskeleton wearer is used as a participant of system control, the wearer can guarantee the front-back walking balance of the whole system in a walking plane, and the exoskeleton is prevented from falling down. Meanwhile, the method utilizes a cascade control strategy to design upper and lower layer controllers, realizes the trajectory planning and trajectory tracking of the under-actuated power-assisted exoskeleton, and is simple to realize, easy to realize in engineering and flexible to control.
Example 2
The embodiment provides an under-actuated support leg assistance exoskeleton which is similar to the under-actuated support leg assistance exoskeleton in embodiment 1, except that a real-time controller in the embodiment directly executes the adaptive robust force control method in embodiment 1, so that each part of the exoskeleton realizes good following and assistance effects on human motion.
Example 3
The embodiment provides an adaptive robust force control device of an under-actuated support leg assisting exoskeleton, which applies the adaptive robust force control method of the under-actuated support leg assisting exoskeleton in embodiment 1. The control device comprises an initialization module I, an initialization module II, an initialization module III, a model establishing module, a reference displacement obtaining module, an actual displacement obtaining module, a conversion module and a following module.
The initialization module is used for initializing the sampling period of the real-time controller, and the sampling period is between 10 and 20 milliseconds. The second initialization module is used for rotating the foot 1 to the horizontal position, rotating the first rod 2, the second rod 6 and the back 10 to the vertical position, initializing the knee joint encoder 3 and the hip joint encoder 7 and zeroing the encoder values. The initialization module is used for initializing the force sensor 11 and zeroing the value of the force sensor 11. These three initialization modules can be combined into an initialization unit that aims to initialize the various parts of the exoskeleton with the corresponding values also zeroed.
The model building module is used for building a physical model of the under-actuated support leg assistance exoskeleton and converting the physical model into a state equation. The physical model comprises a human-computer interface model, a motion model of the exoskeleton mechanical body and a complete constraint model provided by a wearer. The model building module mainly executes step (4) in embodiment 1, and can build each model and state equation.
The reference displacement acquisition module is used for connecting the force sensor 11 with a wearer through a back bandage 12, measuring acting force on the force sensor 11 and acquiring reference displacement of the under-actuated support leg assistance exoskeleton through the upper layer controller. The actual displacement acquisition module is used for acquiring an actual angle value of the under-actuated support leg assisting exoskeleton through the knee joint encoder 3 and the hip joint encoder 7 and acquiring actual displacement of a back 10 contact part according to a positive kinematics model of the exoskeleton system. And the actual displacement acquisition module takes the actual displacement and the reference displacement as input quantities of the lower-layer position tracking controller according to the reference displacement, and the output of the lower-layer position tracking controller is motor driving torque at knee joints and hip joints in the under-actuated support leg assisting exoskeleton. The two displacement acquisition modules are respectively used for executing the steps (5) and (6) in the embodiment 1, and can acquire corresponding displacement information.
The conversion module is used for converting the driving torque of the motor into the control current of the corresponding motor through the drivers of the knee joint motor 4 and the hip joint motor 8. The conversion module mainly executes the step (7) in the embodiment 1 to realize the conversion of the torque and the current. The following module is used for controlling the output torque of the corresponding knee joint motor 4 and hip joint motor 8 through each control current to drive each joint of the under-actuated support leg assisting exoskeleton to rotate, and the following motion of the under-actuated support leg assisting exoskeleton is achieved. The following module mainly executes the step (8) in the embodiment 1 to realize the following movement of the under-actuated support leg assisted exoskeleton.
Example 4
The present embodiments provide a computer terminal comprising a memory, a processor, and a computer program stored on the memory and executable on the processor. And when the processor executes the program, the steps of the adaptive robust force control method of the under-actuated support leg assisted exoskeleton of the embodiment 1 are realized.
When the method in embodiment 1 is applied, the method can be applied in a software form, for example, a program designed to run independently is installed on a computer terminal, and the computer terminal can be a computer, a smart phone, a control system, other internet of things equipment, and the like. The method of embodiment 1 may also be designed as an embedded running program, and installed on a computer terminal, such as a single chip microcomputer.
Example 5
The present embodiment provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the steps of the adaptive robust force control method for the under-actuated support leg assisted exoskeleton of embodiment 1 are realized.
When the method of embodiment 1 is applied, the method may be applied in the form of software, such as a program designed to be independently run by a computer-readable storage medium, which may be a usb disk designed as a usb shield, and the usb disk is designed to be a program for starting the whole method through external triggering.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.

Claims (7)

1. An adaptive robust force control method for an under-actuated support leg assistance exoskeleton, the under-actuated support leg assistance exoskeleton comprising:
a foot portion;
the bottom end of the first rod piece is connected with the foot part;
the bottom end of the second rod piece is rotatably connected with the top end of the first rod piece;
the knee joint motor is used for driving the first rod piece and the second rod piece to rotate relatively;
a knee joint encoder mounted on the knee joint motor;
the back part is rotatably connected with the top end of the second rod piece;
a hip joint motor for driving the back and the second rod to rotate relatively;
a hip joint encoder mounted on the hip joint motor;
a force sensor mounted on the back;
a back strap, both ends of which are connected with the upper end of the back;
a waist band, both ends of which are connected with the lower end of the back part; and
a real-time controller electrically connected to the knee joint motor, the hip joint motor, the knee joint encoder, the hip joint encoder, and the force sensor;
the adaptive robust force control method comprises the following steps:
(1) initializing a sampling period of the real-time controller;
(2) rotating the foot to horizontal and the first bar, the second bar and the back to vertical positions, and initializing the knee joint encoder and the hip joint encoder and zeroing the encoder values;
(3) initializing the force sensor and zeroing the value of the force sensor;
(4) establishing a physical model of the under-actuated support leg assisting exoskeleton and converting the physical model into a state equation; wherein the physical model comprises a human-computer interface model, a motion model of the exoskeleton mechanical body and a complete constraint model provided by a wearer;
(5) the force sensor is connected with a wearer through the back bandage, acting force on the force sensor is measured, and reference displacement of the under-actuated support leg assisting exoskeleton is obtained through the upper layer controller;
(6) obtaining an actual angle value of the under-actuated support leg assistance exoskeleton through the knee joint encoder and the hip joint encoder, and obtaining an actual displacement of the back contact part according to a positive kinematics model of an exoskeleton system; according to the reference displacement obtained in the step (5), taking the actual displacement and the reference displacement as input quantities of a lower-layer position tracking controller, wherein the output of the lower-layer position tracking controller is motor driving torque of knee joints and hip joints in the under-actuated support leg assisting exoskeleton;
(7) converting the motor driving torque obtained in the step (6) into control current of a corresponding motor through the drivers of the knee joint motor and the hip joint motor; and
(8) controlling output torque of the corresponding knee joint motor and hip joint motor through each control current to drive each joint of the under-actuated support leg assisted exoskeleton to rotate, so as to realize follow-up motion of the under-actuated support leg assisted exoskeleton;
wherein the human-machine interface model is:
Figure FDA0003392051290000021
wherein, Fhm＝[Fhmx Fhmy τez]TFor man-machine acting force, T is transposed symbol, x, y and z are symbols representing three-dimensional coordinate axes respectively, and tauezIs moment, Fhmx、FhmyRespectively representing the components of the man-machine acting force in the x and y directions; k ═ diag { K ═ Kx,Ky,KzHuman stiffness of the human interface, Kx,Ky,KzRespectively representing the components of the rigidity of the man-machine interface in the x direction, the y direction and the z direction; x is the number ofh＝[xhx xhy xhz]TAnd xe＝[xex xey xez]TDisplacement of the wearer at the back contact and displacement of the exoskeleton, x, respectivelyhx,xhy,xhzRepresenting the components of the wearer's displacement in the x, y, z directions, x, respectively, at the back contactex,xey,xezRespectively representing the components of the exoskeleton's displacement in the x, y and z directions;
Figure FDA0003392051290000022
model uncertainty and interference are concentrated on the human-machine interface;
by integration of the human-machine forces in the transformation of the physical model
Figure FDA0003392051290000023
In place of FhmObtaining the state equation as follows:
Figure FDA0003392051290000024
the motion model is as follows:
Figure FDA0003392051290000031
in the formula, Fhm＝[Fhmx Fhmy τez]TFor man-machine forces of contact points, τactFor the motor drive torque, J is the Jacobian matrix of the system at the force sensor, q (t) [ q [ q ]) ]1(t),q2(t),q3(t)]TThe rotation angles of the ankle joint, knee joint and hip joint, Msp3(q) is the inertial matrix of the system,
Figure FDA0003392051290000032
is the centrifugal and Coriolis force matrix of the system, Gsp3(q) is the gravity matrix of the system, B is the damping matrix of the system,
Figure FDA0003392051290000033
is the centralized modeling error of the system;
the motion model is further converted into:
Figure FDA0003392051290000034
in the formula (I), the compound is shown in the specification,
Figure FDA0003392051290000035
the complete constraint model is:
xez＝xezd(t)
wherein x isezd(t) is the desired exoskeleton displacement in the z-direction at back contact;
taking the second derivative of the complete constraint model:
Figure FDA0003392051290000036
find out
Figure FDA0003392051290000037
And τezFinally, the following can be obtained:
Figure FDA0003392051290000038
Figure FDA0003392051290000039
in the formula, xea＝[xex xey]T,Mea＝u1Mxu4,Bea＝u1J-Tu3,Cea＝u1Cxu4,Gea＝u1Gx,Bxea＝u1Bxu4,Mea2＝u2Mxu4,Cea2＝u2Cxu4,Gea2＝u2Gx,Bxea2＝u2Bxu4,Bea2＝-u2J-Tu3,
Figure FDA0003392051290000041
Figure FDA0003392051290000042
u2＝[0 0 1],
Figure FDA0003392051290000043
2. The method for adaptive robust force control of an under-actuated support leg assisted exoskeleton of claim 1 wherein the method of converting the physical model to the equation of state comprises the steps of:
(4.1) order State variables
Figure FDA0003392051290000044
Wherein F'hm＝[Fhmx Fhmy]T,
Figure FDA0003392051290000045
Let the centralized model uncertainty be:
Figure FDA0003392051290000046
(4.2) dividing the uncertainty of the centralized model into a constant part and a time-varying function part to obtain
Figure FDA0003392051290000047
ΔinAnd ΔiRespectively represent
Figure FDA0003392051290000048
A constant portion and a time-varying portion of;
(4.3) is provided with
Figure FDA0003392051290000049
Wherein, Kθa＝[1/Kx 1/Ky]T，Δ1an＝[Δ1anx Δ1any]T，Δ1anx、Δ1anyRespectively representing the uncertainty Δ of the lumped model1anComponents in the x, y directions; beta ═ Y2 Y3 Y4 X4 J2 J3 J4]TSystem parameter, Y, for exoskeleton support legs2、Y3、Y4、X4、J2、J3、J4Model parameters, B, both of mechanical structureθ＝[B1 B2 B3]T，B1、B2、B3Respectively representing the damping coefficients of the ankle joint, the knee joint and the hip joint; delta3an＝[Δ3anx Δ3any]T；Δ3anx、Δ3anyAre respectively a setMedium interference delta3anThe components in the x, y directions.
3. The method of adaptive robust force control for an under-actuated support leg assisted exoskeleton of claim 2, wherein the state equation of the physical model of the under-actuated support leg assisted exoskeleton is:
Figure FDA00033920512900000410
Figure FDA00033920512900000411
Figure FDA00033920512900000412
wherein, Kxy＝diag{Kx,Ky}。
4. The adaptive robust force control method for an under-actuated support leg assisted exoskeleton of claim 3 wherein the control method for the upper level controller comprises the steps of:
according to the state equation of the physical model in the step (4), setting the first tracking error as z1a＝x1a-x1adWherein x is1adThe integral of the expected man-machine acting force in the x and y directions is 0; let xmFor the first virtual control input, the first virtual control input xmFirst tracking error z for man-machine effort1aRapidly towards zero;
let xm＝xma+xms+xmsnWherein
Figure FDA0003392051290000051
xms＝K1z1a，fθFAnd YθFIs formed by xmaIs linearized with respect to the parameters of (f)θF＝[0 0]T，
Figure FDA0003392051290000052
xmsnIs a robust feedback term;
Figure FDA0003392051290000053
desired human-machine forces in the x, y directions, respectively, K1＝diag{K1x,K1yIs a linear feedback gain matrix, Kf＝diag{1/Kx,1/Ky}；
Figure FDA0003392051290000054
Respectively represents Kf,Δ1anAn estimated value of (d);
Figure FDA0003392051290000055
represents KθaAn estimated value of (d);
Figure FDA0003392051290000056
to represent
Figure FDA0003392051290000057
Transposing;
Figure FDA0003392051290000058
to represent
Figure FDA0003392051290000059
Transposing;
Figure FDA00033920512900000510
is thetaFAnd the range of the estimated values is:
Figure FDA00033920512900000511
wherein, thetaFmaxi，θFminiAre each thetaFThe maximum and minimum values of the ith element;
Figure FDA00033920512900000512
to a parameter thetaFIs estimated value of
Figure FDA00033920512900000513
The minimum value of (a) is determined,
Figure FDA00033920512900000514
to a parameter thetaFIs estimated value of
Figure FDA00033920512900000515
Maximum value of (d); estimated value
Figure FDA00033920512900000516
In the upper layer controller by adaptive rate
Figure FDA00033920512900000517
Is obtained in which
Figure FDA00033920512900000518
Γ1Is a matrix of positive fixed gains, and,
Figure FDA00033920512900000519
the mapping function of (d) is:
Figure FDA00033920512900000520
in the formula-iIs an independent variable; x is the number ofmsnSatisfies the following conditions:
Figure FDA00033920512900000521
in the formula (I), the compound is shown in the specification,
Figure FDA00033920512900000522
is an estimated value
Figure FDA00033920512900000523
Minus the actual value thetaF，
Figure FDA00033920512900000524
ε1Is a threshold and is arbitrarily non-negative;
according to a first virtual control input xmiThe reference displacement, the reference speed and the reference acceleration of the exoskeleton are obtained by smoothing the reference displacement, the reference speed and the reference acceleration through a third-order filter, wherein i is 1 and 2; wherein, the state equation of the third-order filter is:
Figure FDA0003392051290000061
Figure FDA0003392051290000062
Figure FDA0003392051290000063
i＝1,2
let yiRepresents the exoskeleton reference displacement, let yi(s)＝xmi(s)，xi(1),xi(2),xi(3) Respectively represent the filtered reference displacement, reference velocity and reference acceleration, then yiTo xi(1) The transfer function is:
Figure FDA0003392051290000064
obtaining x by the transfer functionmiConverting into required smooth exoskeleton reference displacement xi(1) (ii) a Wherein, a1,a2,a3Obtained by pole placement.
5. The adaptive robust force control method for an under-actuated support leg assisted exoskeleton of claim 4 wherein the design method for the lower level position tracking controller comprises the steps of:
setting a second tracking error
Figure FDA0003392051290000065
Wherein
Figure FDA0003392051290000066
Defining a conversion equation:
Figure FDA0003392051290000067
wherein, K2Taking an arbitrary non-negative number, z3Is the third tracking error; z is a radical of2And z3Has a transfer function of
Figure FDA0003392051290000068
Let Bxeax3a＝YBBθ，
Figure FDA0003392051290000069
Wherein, K2iRepresents K2β ═ Y, the ith element of (a)2Y3 Y4 X4 J2 J3 J4]T，Bθ＝[B1 B2 B3]TFor damping of the system, B1、B2、B3Respectively representing the damping coefficients of the ankle joint, the knee joint and the hip joint; control of tauactThe determination method comprises the following steps: tau isact＝τacta+τacts+τactsnWherein, τactsnIn order to be robust with respect to the feedback term,
Figure FDA00033920512900000610
Figure FDA00033920512900000611
K3is a linear feedback gain that is a function of,
Figure FDA00033920512900000612
are respectively beta, Bθ，Δ3anIs determined by the estimated value of (c),
Figure FDA00033920512900000613
Figure FDA00033920512900000614
wherein, YBFrom Bxeax3aParameter linearization to obtain f0And Y is composed of
Figure FDA0003392051290000071
The parameters are obtained by linearization, and the parameters are obtained by linearization,
Figure FDA0003392051290000072
to a parameter thetaqIs estimated value of
Figure FDA0003392051290000073
The minimum value of (a) is determined,
Figure FDA0003392051290000074
to a parameter thetaqIs estimated value of
Figure FDA0003392051290000075
Maximum value of (d); thetaqmaxi，θqminiAre each thetaqThe maximum and minimum values of the ith element; estimated value
Figure FDA0003392051290000076
Is controlled by an adaptation rate in the lower layer position tracking controller
Figure FDA0003392051290000077
Is obtained in which
Figure FDA0003392051290000078
Γ2Is a matrix of positive fixed gains, and,
Figure FDA0003392051290000079
the mapping function of (d) is:
Figure FDA00033920512900000710
in the formula-iIs an independent variable;
let phi3＝[-Y -YB I2*2]T，τactsnSatisfies the following conditions:
Figure FDA00033920512900000711
Figure FDA00033920512900000712
in the formula (I), the compound is shown in the specification,
Figure FDA00033920512900000713
is an estimated value
Figure FDA00033920512900000714
Minus the actual value thetaq，
Figure FDA00033920512900000715
ε3Is a threshold value and is an arbitrary non-negative number,
Figure FDA00033920512900000716
6. the adaptive robust force control method for an under-actuated support leg assisted exoskeleton of claim 1, wherein the back is a back plate and the knee joint encoder and the hip joint encoder are joint rotary encoders; the under-actuated support leg assistance exoskeleton further comprises:
the knee joint motor drives the first rod piece and the second rod piece to rotate relatively through the knee joint reducer;
and the hip joint motor drives the back and the second rod piece to rotate relatively through the hip joint reducer.
7. An adaptive robust force control device of an under-actuated support leg assisting exoskeleton, which applies the adaptive robust force control method of the under-actuated support leg assisting exoskeleton as claimed in any one of claims 1 to 6, characterized by comprising the following steps:
the initialization module I is used for initializing a sampling period of the real-time controller, and the sampling period is between 10 and 20 milliseconds;
a second initialization module, configured to rotate the foot to a horizontal position, rotate the first rod, the second rod, and the back to a vertical position, initialize the knee joint encoder and the hip joint encoder, and zero an encoder value;
the initialization module III is used for initializing the force sensor and zeroing the numerical value of the force sensor;
the model establishing module is used for establishing a physical model of the under-actuated support leg assistance exoskeleton and converting the physical model into a state equation; wherein the physical model comprises a human-computer interface model, a motion model of the exoskeleton mechanical body and a complete constraint model provided by a wearer;
the reference displacement acquisition module is used for connecting the force sensor with a wearer through the back bandage, measuring acting force on the force sensor and acquiring reference displacement of the under-actuated support leg assistance exoskeleton through the upper layer controller;
the actual displacement acquisition module is used for acquiring an actual angle value of the under-actuated support leg assistance exoskeleton through the knee joint encoder and the hip joint encoder and acquiring an actual displacement of the back contact part according to a positive kinematics model of an exoskeleton system; the actual displacement acquisition module takes the actual displacement and the reference displacement as input quantities of a lower-layer position tracking controller according to the reference displacement, and the output of the lower-layer position tracking controller is motor driving torque at knee joints and hip joints in the under-actuated support leg assisting exoskeleton;
the conversion module is used for converting the driving torque of the motors into control currents of the corresponding motors through the drivers of the knee joint motor and the hip joint motor; and
and the following module is used for controlling the corresponding output torque of the knee joint motor and the hip joint motor through each control current to drive each joint of the under-actuated support leg assistance exoskeleton to rotate so as to realize the following motion of the under-actuated support leg assistance exoskeleton.
CN202011363672.0A 2020-11-27 2020-11-27 Adaptive robust force control method and device for under-actuated support leg assistance exoskeleton Active CN112388616B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011363672.0A CN112388616B (en) 2020-11-27 2020-11-27 Adaptive robust force control method and device for under-actuated support leg assistance exoskeleton

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011363672.0A CN112388616B (en) 2020-11-27 2020-11-27 Adaptive robust force control method and device for under-actuated support leg assistance exoskeleton

Publications (2)

Publication Number Publication Date
CN112388616A CN112388616A (en) 2021-02-23
CN112388616B true CN112388616B (en) 2022-02-22

Family

ID=74604731

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011363672.0A Active CN112388616B (en) 2020-11-27 2020-11-27 Adaptive robust force control method and device for under-actuated support leg assistance exoskeleton

Country Status (1)

Country Link
CN (1) CN112388616B (en)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105116728A (en) * 2015-08-14 2015-12-02 上海申磬产业有限公司 A method for single-joint power-assisting exoskeleton inversion self-adaptive robust force control
CN105137972A (en) * 2015-08-14 2015-12-09 浙江大学 Adaptive robustness cascading force controlling method for single-joint powered exoskeleton
CN107137207A (en) * 2017-07-03 2017-09-08 哈尔滨工业大学 Drive lacking lower limb assistance exoskeleton robot based on rope pulley mechanism
KR20170130768A (en) * 2016-05-19 2017-11-29 현대자동차주식회사 System and methode for determining wearer's intention of exoskeleton robot
KR20180085983A (en) * 2017-01-20 2018-07-30 국방과학연구소 Downhill walking control method of wearable exoskeleton robot
CN109129464A (en) * 2018-07-06 2019-01-04 合肥工业大学 The restraining force robust method of servo-controlling and device of drive lacking flexible mechanical arm system
CN109746902A (en) * 2019-03-13 2019-05-14 北京理工大学 A kind of upper limb ectoskeleton cooperative control method with dynamic load compensation

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105116728A (en) * 2015-08-14 2015-12-02 上海申磬产业有限公司 A method for single-joint power-assisting exoskeleton inversion self-adaptive robust force control
CN105137972A (en) * 2015-08-14 2015-12-09 浙江大学 Adaptive robustness cascading force controlling method for single-joint powered exoskeleton
KR20170130768A (en) * 2016-05-19 2017-11-29 현대자동차주식회사 System and methode for determining wearer's intention of exoskeleton robot
KR20180085983A (en) * 2017-01-20 2018-07-30 국방과학연구소 Downhill walking control method of wearable exoskeleton robot
CN107137207A (en) * 2017-07-03 2017-09-08 哈尔滨工业大学 Drive lacking lower limb assistance exoskeleton robot based on rope pulley mechanism
CN109129464A (en) * 2018-07-06 2019-01-04 合肥工业大学 The restraining force robust method of servo-controlling and device of drive lacking flexible mechanical arm system
CN109746902A (en) * 2019-03-13 2019-05-14 北京理工大学 A kind of upper limb ectoskeleton cooperative control method with dynamic load compensation

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
Precision Cascade Force Control of Multi-DOF Hydraulic Leg Exoskeleton;Shan Chen;《IEEE Access》;20180312;第6卷;第8574-8583页 *

Also Published As

Publication number Publication date
CN112388616A (en) 2021-02-23

Similar Documents

Publication Publication Date Title
CN112775937B (en) Self-adaptive robust control method and device for under-actuated hydraulic single-leg power-assisted exoskeleton
JP3443077B2 (en) Robot motion pattern generation device and motion pattern generation method, and robot
Chevallereau et al. Optimal reference trajectories for walking and running of a biped robot
US7366587B2 (en) Legged mobile robot
US8556836B2 (en) System for controlling an exoskeleton haptic device for rehabilitation purposes, and corresponding exoskeleton haptic device
JP5052013B2 (en) Robot apparatus and control method thereof
US8271137B2 (en) Robot and method of controlling the same
CN110405762B (en) Biped robot attitude control method based on spatial second-order inverted pendulum model
CN113001540B (en) Man-machine interaction intelligent control method of load-mobile exoskeleton and exoskeleton system
CN112051741A (en) Dynamic motion generation and control method for biped robot
JP2007210071A (en) Robot device and its control method
WO2004030870A1 (en) Robot device and control method of robot device
WO2003078110A1 (en) Robot device, legged locomotion robot operation control device and operation control method, legged locomotion robot sensor system, and locomotion device
WO2002045913A1 (en) Device and method for controlling operation of legged robot, and robot device
CN108897220B (en) Self-adaptive stable balance control method and system and biped humanoid robot
WO2002040222A1 (en) Device and method for controlling motion of legged mobile robot, and motion unit generating method for legged mobile robot
CN114397810A (en) Four-legged robot motion control method based on adaptive virtual model control
WO2004033160A1 (en) Robot device operation control device and operation control method
Zhu et al. Design and control of BRAVER: a bipedal robot actuated via proprioceptive electric motors
Kim et al. Realization of dynamic stair climbing for biped humanoid robot using force/torque sensors
CN112388616B (en) Adaptive robust force control method and device for under-actuated support leg assistance exoskeleton
JP5616289B2 (en) Floor surface estimation device
JP2003117858A (en) Method and device for control of robot walk
CN110308727A (en) A kind of control method for eliminating biped robot's upper body posture shaking
CN116000917A (en) Motion trail safety triggering data driving control method of rehabilitation walking robot

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant