A decentralized fault-tolerant control method for flexible joint space robots without velocity feedback

Abstract

The invention discloses a decentralized fault-tolerant control method for a flexible joint space robot without speed feedback, comprising the following steps: S1: constructing a dynamic model of a flexible joint space robot system whose carrier posture is not controlled; S2: introducing a flexible compensator; based on a singular perturbation method, decomposing the model into a fast-changing subsystem representing joint flexible vibration and a slow-changing subsystem representing arm trajectory tracking; S3: for the fast-changing subsystem, designing a torque differential feedback controller; S4: for the slow-changing subsystem, using a speed observer to identify the joint angular velocity, and using a decentralized neural network to identify the uncertainty and fault terms of the model, and designing a decentralized fault-tolerant control method without speed feedback accordingly; the beneficial technical effects of the invention are: solving the problem that the singular perturbation method cannot be applied to a robot system with a small joint stiffness, using a decentralized neural network to reconstruct the model to improve the task execution efficiency of the space robot, and the robot arm can stably track the expected trajectory.

Description

Technical Field

[0001] The invention belongs to the technical field of mechanical arm control, and in particular relates to a speed-feedback-free decentralized fault-tolerant control method for a flexible joint space robot. Background Art

[0002] In recent years, with the continuous and vigorous development of aerospace technology, advanced spacecraft represented by space robots have received increasing attention. Space robots can assist or replace astronauts in performing complex and high-risk extravehicular missions in a space environment with extreme temperature differences, ultra-strong radiation, and near-weightlessness, and play an unparalleled role in improving work efficiency and ensuring the safety of astronauts. Considering the limitation of the operation window time, the space robot needs to quickly capture and drag the target object within a fixed time, but due to the transmission effect of the gear harmonic reducer inside the mechanical arm joint, the rotation of the mechanical arm connecting rod and the rotation of the joint motor cannot be synchronized, that is, there is an angle deviation between the two, which causes the flexible vibration of the joint hinge. At present, from the existing data, the research on the dynamics and vibration reduction control of flexible arm space robots is relatively mature, and the singular perturbation method is usually used, which cannot be applied to robot systems with small joint stiffness.

[0003] It should be pointed out that due to the influence of space charged particle jets and solar magnetic fields, the space robot's own sensor instruments and actuators are prone to unknown failures, resulting in model uncertainty and the inability to obtain speed signals, making it impossible to stably track the desired trajectory. However, there is no report on the fault-tolerant control of space robots that considers the above two types of failures. In view of this, the present invention designs a speed-free feedback-free decentralized fault-tolerant control method for a flexible joint space robot with the help of a speed observer and a decentralized neural network. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a decentralized fault-tolerant control method for a flexible joint space robot without speed feedback, which is used to solve the technical problems of sensor and actuator failures, model uncertainty and inability to obtain speed signals.

[0005] To achieve the above object, the present invention provides a decentralized fault-tolerant control method for a flexible joint space robot without speed feedback, comprising the following steps:

[0006] S1: Combining the momentum conservation relation with the Lagrangian method, a dynamic model of a flexible joint space robot system with uncontrolled carrier posture is constructed;

[0007] S2: Introduce a flexible compensator to improve the equivalent stiffness of the model; based on the singular perturbation method, decompose the model into a fast-changing subsystem representing the flexible vibration of the joint and a slow-changing subsystem representing the trajectory tracking of the arm;

[0008] S3: For fast-changing subsystems, design a torque differential feedback controller;

[0009] S4: For the slow-varying subsystem, the speed observer is used to identify the joint angular velocity, and the uncertainties and faults of the model are identified with the help of a decentralized neural network. Based on this, a decentralized fault-tolerant control method without speed feedback is designed.

[0010] Further: the specific implementation process of S1 is:

[0011] Combining the conservation relations of the system's angular momentum and linear momentum with the Lagrangian method, the fully driven dynamic equation of the flexible joint space robot system with uncontrolled carrier posture is as follows:

[0012] (1)

[0013] (2)

[0014] (3)

[0015] in, is the symmetric and positive definite inertia matrix of the system; is the column vector of the system including Coriolis force and centrifugal force; is the column vector of control torque transmitted by the flexible joint; is the control torque column vector of the joint drive motor; ; is the robot arm rotation angle column vector; is the driving motor rotor angle column vector; is the positive inertia matrix of the driving motor; is the joint stiffness coefficient matrix.

[0016] Further: the specific implementation process of S2 is:

[0017] For the dynamic model of the flexible joint space robot system determined by S1, a singular perturbation method is used to design a controller;

[0018] First, the dynamic model of the flexible joint space robot system determined by S1 can be obtained;

[0019] (4)

[0020] in, is the positive inertia matrix of the driving motor;

[0021] is the joint stiffness coefficient matrix;

[0022] is the control torque column vector of the joint drive motor;

[0023] is the column vector of the control torque transmitted by the flexible joint, express The second derivative with respect to time t;

[0024] is the column vector of the space robot arm’s rotation angle, express The second derivative with respect to time t;

[0025] The control law of the drive motor is designed as:

[0026] (5)

[0027] in, For the newly introduced joint flexibility compensation controller, is a diagonal, positive definite flexible compensation matrix; , is the identity matrix; is the control input to be designed;

[0028] Arranging the above formulas (5) and (4) yields:

[0029] (6)

[0030] in, is the equivalent stiffness matrix after flexibility compensation. You can change ;

[0031] The dynamic model of the flexible joint space robot system determined by S1 and formula (6) constitute the dynamic model of the flexible joint space robot after flexibility compensation;

[0032] Combined with the singular perturbation method, we can Broken down into two parts, as follows:

[0033] (7)

[0034] in, and They represent the fast-changing and slow-changing subsystem control laws of the flexible joint space robot respectively;

[0035] Assuming the stiffness coefficient matrix The minimum stiffness is , define the singular perturbation factor for: ,make ,but is a diagonal, positive definite constant matrix of the same order of magnitude as the slowly varying component;

[0036] Combining equation (6) and equation (7), the dynamic equation of the fast-changing subsystem can be derived as:

[0037] (8)

[0038] At this time, if ,but and .

[0039] Further: the slow-changing subsystem in S4 is specifically: Substituting the fast-changing subsystem in S3 into the dynamics model of the flexible joint space robot system determined in S1, the dynamics equation of the slow-changing subsystem of the flexible joint space robot can be obtained as follows:

[0040] (9)

[0041] in, is the positive definite inertia matrix; To use Simplify and organize The resulting column vector;

[0042] express The second derivative with respect to time t.

[0043] Furthermore: the specific implementation process of S3 is:

[0044] For the fast-changing subsystem in S2, in order to suppress the elastic vibration caused by the flexible joint, the torque differential feedback controller is designed as:

[0045] (10)

[0046] in, is the diagonal coefficient matrix;

[0047] is the defined singular perturbation factor;

[0048] is the column vector of the control torque transmitted by the flexible joint, express The first derivative with respect to time t.

[0049] Further: the specific implementation process of S4 is:

[0050] In order to achieve individual control of joint motion, each joint module is considered as a second-order subsystem, and the local variables are separated from the slow-varying subsystem. , then the decentralized dynamics model of the slow-varying subsystem can be expressed as:

[0051] (11)

[0052] (12)

[0053] in, is the inertia parameter term of the second-order subsystem; are the Coriolis force and centrifugal force terms of the second-order subsystem; , , and Respectively, vector , , and The i-th element of ; and The matrices are and the ijth element of ; is the second-order subsystem cross-linking term;

[0054] For a space robot with a joint actuator failure, the decentralized dynamics equation of its slow-varying subsystem can be described as:

[0055] (13)

[0056] in, represents the failure factor, satisfying , the smaller the value, the more serious the fault of the corresponding actuator;

[0057] Define the state vector of the second-order subsystem , then equation (13) can be written as the following state space equation:

[0058] (14)

[0059] in, ;

[0060] ;

[0061] ;

[0062] The speed observer of the second-order subsystem is designed as:

[0063] (15)

[0064] in, and They are and The estimated value of is a saturation function;

[0065] When the desired trajectory of the system joints , and exists and is bounded;

[0066] The position error vector of the second-order subsystem is defined as:

[0067] (16)

[0068] The position loop integral sliding mode function of the second-order subsystem is designed as:

[0069] (17)

[0070] in, is a positive constant;

[0071] Taking the derivative of equation (17) with respect to time t, we can obtain:

[0072] (18)

[0073] when , there exists a small positive number ,satisfy:

[0074] (19)

[0075] Substituting formula (19) into formula (18), we can get

[0076] (20)

[0077] Define the virtual angular velocity of the second-order subsystem as

[0078] (twenty one)

[0079] in, is a normal number, is a symbolic function;

[0080] The virtual angular velocity tracking error is defined as:

[0081] (twenty two)

[0082] The velocity loop integral sliding mode function of the second-order subsystem is designed as:

[0083] (twenty three)

[0084] in, is a positive constant;

[0085] Taking the derivative of equation (23) with respect to time t, we can obtain:

[0086] (twenty four)

[0087] Using RBF (Radial Basis Function) neural network to analyze the uncertainty of the second-order subsystem and unknown items Reconstruct, assuming that the ideal neural network function expression is as follows:

[0088] (25)

[0089] (26)

[0090] in, and is the ideal neural network weight;

[0091] and is the neural network basis function;

[0092] and is the neural network approximation error, is an unknown positive constant;

[0093] definition and They are and The estimated value of and is the Gaussian basis function of the neural network; then and The actual RBF neural network outputs are:

[0094] (27)

[0095] (28)

[0096] Define the neural network weight estimation error as , , the basis function estimation error is , , combined with equations (25) to (28), we have:

[0097] (29)

[0098] (30)

[0099] When the cross-link term of the second-order subsystem is bounded and satisfies:

[0100] (31)

[0101] in, , ;

[0102] Using the neural network of formula (32) Compensation cross-link Impact on the second-order subsystem:

[0103] (32)

[0104] in, is the neural network expression, is the ideal neural network weight The estimated value of is the neural network basis function An estimated value of

[0105] The minimum approximation error of the neural network for the second-order subsystem is defined as follows:

[0106] (33)

[0107] (34)

[0108] (35)

[0109] in, .

[0110] Neural network weights , and Update according to the following adaptive laws:

[0111] (36)

[0112] (37)

[0113] (38)

[0114] The speed loop integral sliding mode fault-tolerant control law of the second-order subsystem is designed as:

[0115] (39)

[0116] in, and All are normal numbers.

[0117] The beneficial technical effects of the present invention are:

[0118] 1. Based on the introduction of the singular perturbation method, a flexible compensator is designed to improve the equivalent stiffness of the joint and solve the problem that the singular perturbation method cannot be applied to robot systems with small joint stiffness.

[0119] 2. This invention designs a speed-free feedback-free decentralized fault-tolerant control method for the slow-changing subsystem of a flexible joint space robot, which solves the motion control problem of the space robot under concurrent failures of sensors and actuators. The velocity observer is used to identify the joint angular velocity, and the model uncertainty and fault terms are reconstructed with the help of a decentralized neural network, which comprehensively improves the task execution efficiency of the space robot.

[0120] 3. The present invention designs a torque differential feedback controller for the fast-changing subsystem of the flexible joint space robot to suppress the elastic vibration caused by the flexible joint and ensure that the robot arm can stably track the desired trajectory. BRIEF DESCRIPTION OF THE DRAWINGS

[0121] In order to more clearly illustrate the embodiments of the present invention and the technical solutions in the prior art, the drawings required for use in the embodiments and the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some implementation methods recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0122] Figure 1 It is a principle block diagram of the present invention;

[0123] Figure 2 is a plan view of the flexible joint space robot in the present invention;

[0124] Figure 3 is the angular displacement curve of the space robot joint 1 of the present invention (flexible compensator closed);

[0125] Figure 4 is the angular displacement curve of the space robot joint 2 of the present invention (flexible compensator closed);

[0126] Figure 5 is the angular displacement curve of the space robot joint 1 of the present invention (flexible compensator opened);

[0127] Figure 6 is the angular displacement curve of the space robot joint 2 of the present invention (flexible compensator opened);

[0128] Figure 7 The angular velocity observation curve of the space robot joint 1 of the present invention (flexible compensator closed or opened);

[0129] Figure 8 This is the angular velocity observation curve of the space robot joint 2 of the present invention (flexible compensator closed or opened). DETAILED DESCRIPTION

[0130] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments provided in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0131] Obviously, the drawings described below are only some examples or embodiments of the present application. For ordinary technicians in this field, the present application can also be applied to other similar scenarios based on these drawings without creative work. In addition, it can also be understood that although the efforts made in this development process may be complicated and lengthy, for ordinary technicians in this field related to the content disclosed in this application, some changes in design, manufacturing or production based on the technical content disclosed in this application are just conventional technical means, and should not be understood as insufficient content disclosed in this application.

[0132] Unless otherwise specified, all embodiments and optional embodiments of the present application can be combined with each other to form a new technical solution.

[0133] See also Figure 1 , is a principle block diagram of the present invention, comprising the following steps:

[0134] S1: Construction of the dynamic model of the flexible joint space robot system based on the angular momentum conservation relationship and the Lagrangian method, including:

[0135] Combining the conservation relations of the system's angular momentum and linear momentum with the Lagrange equation, the fully driven dynamic equation of the flexible joint space robot system with uncontrolled carrier posture is obtained as follows:

[0136] (1)

[0137] (2)

[0138] (3)

[0139] in, is the symmetric and positive definite inertia matrix of the system; is the column vector of the system including Coriolis force and centrifugal force; is the column vector of control torque transmitted by the flexible joint; is the control torque column vector of the joint drive motor; ; is the robot arm rotation angle column vector; is the driving motor rotor angle column vector; is the positive inertia matrix of the driving motor; is the joint stiffness coefficient matrix.

[0140] Step S2: Deconstruction of the flexible joint space robot system model integrating the flexible compensator and singular perturbation technology, specifically including:

[0141] For the flexible joint space robot dynamics model determined by equations (1)-(3), the present invention adopts a singular perturbation method to design a controller.

[0142] It is worth noting that the traditional singular perturbation method is only applicable to robot systems with large joint stiffness. For space robots with small joint stiffness, its application range will be greatly limited. In order to solve this problem, a flexible compensator is introduced to increase the equivalent stiffness of the joint, ensuring that the traditional singular perturbation method can be used for the motion control of space robots with strong joint flexibility.

[0143] From formula (2) and formula (3), we can get:

[0144] (4)

[0145] in, is the positive inertia matrix of the driving motor;

[0146] is the joint stiffness coefficient matrix;

[0147] is the control torque column vector of the joint drive motor;

[0148] is the column vector of the control torque transmitted by the flexible joint, express The second derivative with respect to time t;

[0149] is the column vector of the space robot arm’s rotation angle, express The second derivative with respect to time t;

[0150] The control law of the drive motor is designed as:

[0151] (5)

[0152] in, For the newly introduced joint flexibility compensation controller, is a diagonal, positive definite flexible compensation matrix; , is the identity matrix; is the control input to be designed.

[0153] Substituting formula (5) into formula (4), we can get:

[0154] (6)

[0155] in, is the equivalent stiffness matrix after flexibility compensation. You can change .

[0156] Equation (1) and equation (6) constitute the dynamic model of the flexible joint space robot after flexibility compensation.

[0157] Combined with the singular perturbation method, we can Broken down into two parts, as follows:

[0158] (7)

[0159] in, and They represent the fast-changing and slow-changing subsystem control laws of the flexible joint space robot respectively;

[0160] Assuming the stiffness coefficient matrix The minimum stiffness is , define the singular perturbation factor for: ,make ,but is a diagonal, positive definite constant matrix of the same order of magnitude as the slowly varying component.

[0161] Combining equation (6) and equation (7), the dynamic equation of the fast-changing subsystem can be derived as:

[0162] (8)

[0163] At this time, if ,but and

[0164] Substituting equation (8) into equation (1), the dynamic equation of the slow-varying subsystem of the flexible joint space robot can be obtained as follows:

[0165] (9)

[0166] in, is the positive definite inertia matrix; To use Simplify and organize The resulting column vector.

[0167] Step S3: Design of the controller of the fast-changing subsystem of the flexible joint space robot, specifically including:

[0168] For the fast-changing subsystem (8), in order to suppress the elastic vibration caused by the flexible joint, the torque differential feedback controller is designed as

[0169] (10)

[0170] in, is the diagonal coefficient matrix;

[0171] is the defined singular perturbation factor;

[0172] is the column vector of the control torque transmitted by the flexible joint, express The first derivative with respect to time t.

[0173] Step S4: Design of the controller of the slow-changing subsystem of the flexible joint space robot, specifically including:

[0174] The control objective of the slow-varying subsystem is to design a slow-varying controller to compensate for the influence of model unknowns and actuator failure on the system control performance, so as to achieve stable tracking of the desired trajectory by the joint links of the space robot without speed signal feedback.

[0175] In order to achieve independent control of joint motion, each joint module is considered as a second-order subsystem, and the local variables are separated from the slow-varying subsystem (9): , then the decentralized dynamics model of the slow-varying subsystem can be expressed as:

[0176] (11)

[0177] (12)

[0178] in, is the inertia parameter term of the second-order subsystem; are the Coriolis force and centrifugal force terms of the second-order subsystem; , , and Respectively, vector , , and The i-th element of ; and The matrices are and the ijth element of ; is the second-order subsystem cross-linking term.

[0179] For a space robot with a joint actuator failure, the decentralized dynamics equation of its slow-varying subsystem can be described as:

[0180] (13)

[0181] in, represents the failure factor, satisfying , the smaller the value, the more serious the fault of the corresponding actuator;

[0182] Define the state vector of the second-order subsystem , then equation (13) can be written as the following state space equation:

[0183] (14)

[0184] in, ;

[0185] ;

[0186] ;

[0187] The speed observer of the second-order subsystem is designed as:

[0188] (15)

[0189] in, and They are and The estimated value of is a saturation function;

[0190] When the desired trajectory of the system joints , and exists and is bounded;

[0191] The position error vector of the second-order subsystem is defined as:

[0192] (16)

[0193] The position loop integral sliding mode function of the second-order subsystem is designed as:

[0194] (17)

[0195] in, Is a normal number.

[0196] Taking the derivative of equation (17) with respect to time t, we can obtain:

[0197] (18)

[0198] when , there exists a small positive number ,satisfy:

[0199] (19)

[0200] Substituting formula (19) into formula (18), we can get

[0201] (20)

[0202] Define the virtual angular velocity of the second-order subsystem as

[0203] (twenty one)

[0204] in, is a normal number, is a symbolic function;

[0205] The virtual angular velocity tracking error is defined as:

[0206] (twenty two)

[0207] The velocity loop integral sliding mode function of the second-order subsystem is designed as:

[0208] (twenty three)

[0209] in, is a positive constant;

[0210] Taking the derivative of equation (23) with respect to time t, we can obtain:

[0211] (twenty four)

[0212] Using RBF (Radial Basis Function) neural network to analyze the uncertainty of the second-order subsystem and unknown items Reconstruct, assuming that the ideal neural network function expression is as follows:

[0213] (25)

[0214] (26)

[0215] in, and is the ideal neural network weight;

[0216] and is the neural network basis function;

[0217] and is the neural network approximation error, is an unknown positive constant;

[0218] definition and They are and The estimated value of and is the Gaussian basis function of the neural network; then and The actual RBF neural network outputs are:

[0219] (27)

[0220] (28)

[0221] Define the neural network weight estimation error as , , the basis function estimation error is , , combined with equations (25) to (28), we have:

[0222] (29)

[0223] (30)

[0224] When the cross-link term of the second-order subsystem is bounded and satisfies:

[0225] (31)

[0226] in, , ;

[0227] Using the neural network of formula (32) Compensation cross-link Impact on the second-order subsystem:

[0228] (32)

[0229] in, is the neural network expression, is the ideal neural network weight The estimated value of is the neural network basis function An estimated value of

[0230] The minimum approximation error of the neural network for the second-order subsystem is defined as follows:

[0231] (33)

[0232] (34)

[0233] (35)

[0234] in, .

[0235] Neural network weights , and Update according to the following adaptive laws:

[0236] (36)

[0237] (37)

[0238] (38)

[0239] The speed loop integral sliding mode fault-tolerant control law of the second-order subsystem is designed as:

[0240] (39)

[0241] in, and All are normal numbers.

[0242] Theorem 1: For the second-order subsystem dynamics model of the space robot with efficiency loss failure of the actuator (Eq. (14)), under the premise that Assumptions 1 and 2 are valid, the decentralized fault-tolerant control law of Eq. (39) and the adaptive update law shown in Eqs. (36)-(38) are designed to make the trajectory tracking error of the joint converge asymptotically to zero.

[0243] Proof: Select the Lyapunov function as:

[0244] (40)

[0245] in, , .

[0246] The conclusion of Theorem 1 is proved through the following two steps.

[0247] Step 1: Taking the derivative with respect to time t, we get

[0248] (41)

[0249] Substituting equation (24) into equation (41), we obtain:

[0250] (42)

[0251] Substituting equation (39) into equation (42), we can obtain

[0252] (43)

[0253] Substituting equation (31) into equation (43), we obtain:

[0254] (44)

[0255] Notice , using Chebyshev inequality, we can get:

[0256] (45)

[0257] Combining equations (44) and (45), we can obtain:

[0258] (46)

[0259] Obviously, if you choose satisfy ,but .

[0260] Step 2: Taking the derivative with respect to time t, we can obtain:

[0261] (47)

[0262] Obviously, if you choose satisfy ,but .

[0263] From the above analysis, we can see that reasonable selection of parameters can satisfy According to Lyapunov stability theorem and Barbalat lemma, , ,again and They are and function, so the system state error It will also converge to zero asymptotically; Theorem 1 is proved.

[0264] In summary, designing a torque differential feedback controller and a speed-free feedback decentralized fault-tolerant controller for the fast-changing subsystem and slow-changing subsystem of the flexible joint space robot respectively can eliminate the elastic vibration of the joint and ensure the stable tracking control of the space robot's joint links relative to the desired trajectory when the actuator and sensor fail concurrently.

[0265] See also Figure 2 , is a plan view of the flexible joint space robot in the present invention, specifically a schematic diagram of the flexible joint space robot under the influence of concurrent failures of sensors and actuators, the system consists of a free-floating carrier , Rigid rod composition, and Flexible joints are used to connect them. The symbols in the figure are defined as follows: is the inertial coordinate system of the system, For split The main axis coordinate system of the system is In-plane motion, For connecting rod The axis of symmetry; is the total center of mass of the system, is the center of mass of the carrier, and Rotation center of the carrier coincide, For connecting rod The centroid of For each part Relative to the origin of inertial coordinates Radius of The quality is , the moment of inertia is ; For connection and The center of rotation of the hinge; is the carrier attitude angle, For connecting rod The joint angle; and The distance between , and The distance between , The length is Flexible joints The moment of inertia of the motor rotor is , the turning angle is , stiffness coefficient .

[0266] See also Figure 3-Figure 8, is the change of the angular displacement curve and the angular velocity observation curve of a specific embodiment of the present invention under different conditions, which is specifically obtained by the following example;

[0267] by Figure 2 Taking the flexible joint space robot shown in the figure as an example, a program is written in MATLAB to simulate and verify the feasibility of the speed feedback-free decentralized fault-tolerant control method proposed in the present invention.

[0268] The physical parameters of the system are shown in Table 1. The expected trajectory of the space robot joint (unit: radian) is , , the initial position of the system (unit: radians) is , ;

[0269] .

[0270] The controller parameter settings are shown in Table 2:

[0271] .

[0272] At the initial moment of the simulation, the space robot system has actuator failure and sensor failure. The output torque loss rates of joint actuator 1 and joint actuator 2 are 60% and 30% respectively. The sensor cannot directly measure or indirectly obtain the angular velocity signal of the joint.

[0273] Figure 3 and Figure 4 They are the angular displacement curves of joint 1 and joint 2 of the space robot when the flexible compensator is closed, where the dotted line represents the expected trajectory and the solid line represents the actual trajectory;

[0274] Figure 5 and Figure 6 They are the angular displacement curves of joint 1 and joint 2 of the space robot when the flexible compensator is turned on, where the dotted line represents the expected trajectory and the solid line represents the actual trajectory;

[0275] Figure 7 and Figure 8 They are the angular velocity observation curves of joint 1 and joint 2 of the space robot when the flexible compensator is closed, where the dotted line represents the actual angular velocity and the solid line represents the observed angular velocity.

[0276] Depend on Figure 3-Figure 6 It can be seen that the hybrid fault-tolerant control scheme designed by the present invention that integrates joint flexible compensators and singular perturbation technology can achieve stable tracking control of the joints of the space robot under actuator failure; if the flexible compensator is not turned on, the joints of the space robot will immediately deviate from the desired trajectory and fail to achieve the expected movement.

[0277] Depend on Figure 7 and Figure 8 It can be seen that the speed observer designed in the present invention can realize accurate identification of joint angular velocity signals and eliminate the influence of sensor failure on system control performance.

[0278] The above is a very detailed description of one or more embodiments of the present invention, but the content is only a specific example of the present invention and cannot be considered to limit the scope of application of the present invention. Any other methods and changes proposed based on the content of the present invention should fall within the scope of patent protection of the present invention.

Claims

1. A decentralized fault-tolerant control method for a flexible joint space robot without velocity feedback, characterized in that: The following steps are involved: S1: Combining the momentum conservation relation with the Lagrangian method, a dynamic model of a flexible joint space robot system with uncontrolled carrier posture is constructed; S2: Introduce a flexible compensator to improve the equivalent stiffness of the model; based on the singular perturbation method, decompose the model into a fast-changing subsystem representing the flexible vibration of the joint and a slow-changing subsystem representing the trajectory tracking of the arm; S3: For fast-changing subsystems, design a torque differential feedback controller; S4: For the slow-varying subsystem, the velocity observer is used to identify the joint angular velocity, and the uncertainty and fault terms of the model are identified with the help of a decentralized neural network. Based on this, a decentralized fault-tolerant control method without velocity feedback is designed. The specific implementation process of S1 is as follows: Combining the conservation relations of the system's angular momentum and linear momentum with the Lagrangian method, the fully driven dynamic equation of the flexible joint space robot system with uncontrolled carrier posture is as follows: ； ； ； in, is the symmetric and positive definite inertia matrix of the system; is the column vector of the system including Coriolis force and centrifugal force; is the column vector of control torque transmitted by the flexible joint; is the control torque column vector of the joint drive motor; ; is the robot arm rotation angle column vector; is the driving motor rotor angle column vector; is the positive inertia matrix of the driving motor; is the joint stiffness coefficient matrix; The specific implementation process of S4 is as follows: In order to achieve individual control of joint motion, each joint module is considered as a second-order subsystem, and the local variables are separated from the slow-varying subsystem. , then the decentralized dynamics model of the slow-varying subsystem can be expressed as: ； ； in, is the inertia parameter term of the second-order subsystem; are the Coriolis force and centrifugal force terms of the second-order subsystem; , , and Respectively, vector , , and The i-th element of ; and The matrices are The ii-th and ij-th elements of For vector The i-th element of ; is the second-order subsystem cross-linking term; For a space robot with a joint actuator failure, the decentralized dynamics equation of its slow-varying subsystem can be described as: ； in, represents the failure factor, satisfying , the smaller the value, the more serious the fault of the corresponding actuator; Define the state vector of the second-order subsystem , then the decentralized dynamics equation of the slow-varying subsystem can be written as the following state space equation: ； in, ; ； ； The speed observer of the second-order subsystem is designed as: ； in, and They are and The estimated value of is a saturation function, and All are controller parameters , The corresponding gains respectively; When the desired trajectory of the system joints , and exists and is bounded; The position error vector of the second-order subsystem is defined as: ； The position loop integral sliding mode function of the second-order subsystem is designed as: ； in, is a positive constant, is a controller parameter; is the position error vector of the second-order subsystem; Deriving the position loop integral sliding mode function of the second-order subsystem with respect to time t, we can obtain: ； when , there exists a small positive number ,satisfy: ； Further solving, we can get ； The virtual angular velocity of the second-order subsystem is defined as: ； in, is a normal number, is a symbolic function; The virtual angular velocity tracking error is defined as: ； The velocity loop integral sliding mode function of the second-order subsystem is designed as: ； in, is a positive constant; Taking the derivative of the velocity loop integral sliding mode function of the second-order subsystem with respect to time t, we can obtain: ； in, is the virtual angular velocity of the second-order subsystem The first derivative with respect to time t; Using RBF neural network to analyze the uncertainty of the second-order subsystem and unknown items Reconstruct, assuming that the ideal neural network function expression is as follows: ； ； in, and is the ideal neural network weight; and is the neural network basis function; and is the neural network approximation error, and is an unknown positive constant; definition and They are and The estimated value of and is the Gaussian basis function of the neural network; then and The actual RBF neural network outputs are: ； ； Define the neural network weight estimation error as , , the basis function estimation error is , , combining the ideal neural network function with the actual RBF neural network output, we have: ； ； When the cross-link term of the second-order subsystem is bounded and satisfies: ； in, , ; Using Neural Networks Compensation cross-link The impact on the second-order subsystem is shown in the following formula: ； in, is the neural network expression, is the ideal neural network weight The estimated value of is the neural network basis function An estimated value of The minimum approximation error of the neural network for the second-order subsystem is defined as follows: ； ； ； in, ; n = 2; Neural network weights , and Update according to the following adaptive laws: ； ； ； in, , , All are controller parameters; The speed loop integral sliding mode fault-tolerant control law of the second-order subsystem is designed as: ； in, and are controller parameters and are positive numbers.

2. The method for decentralized fault-tolerant control of a flexible joint space robot without velocity feedback according to claim 1, characterized in that: The specific implementation process of S2 is: For the dynamic model of the flexible joint space robot system determined by S1, a controller is designed by using a singular perturbation method; First, the dynamic model of the flexible joint space robot system determined by S1 can be obtained; ； in, is the positive inertia matrix of the driving motor; is the joint stiffness coefficient matrix; is the control torque column vector of the joint drive motor; is the column vector of the control torque transmitted by the flexible joint, express Relative to time The second derivative of is the column vector of the space robot arm rotation angle, express Relative to time The second derivative of The control law of the drive motor is designed as: ; in, For the newly introduced joint flexibility compensation controller, is a diagonal, positive definite flexible compensation matrix; , is the identity matrix; is the control input to be designed; The control law of the above drive motor and the dynamic model of the flexible joint space robot system can be sorted out as follows: ； in, is the equivalent stiffness matrix after flexibility compensation. You can change ; The dynamic model of the flexible joint space robot system determined by S1 and the above formula constitute the dynamic model of the flexible joint space robot after flexibility compensation; Combined with the singular perturbation method, we can Broken down into two parts, as follows: ； in, and They represent the fast-changing and slow-changing subsystem control laws of the flexible joint space robot respectively; Assuming the stiffness coefficient matrix The minimum stiffness is , define the singular perturbation factor for: ,make ,but is a diagonal, positive definite constant matrix of the same order of magnitude as the slowly varying component; The formula after combining the control law of the drive motor and the dynamic model of the flexible joint space robot system is The corresponding formula can be used to derive the dynamic equation of the fast-changing subsystem: ； At this time, if ,but and .

3. The method for decentralized fault-tolerant control of a flexible joint space robot without velocity feedback according to claim 2 is characterized in that: The slow-changing subsystem in S4 is specifically: Substituting the fast-changing subsystem into the dynamic model of the flexible joint space robot system determined in S1, the dynamic equation of the slow-changing subsystem of the flexible joint space robot can be obtained as follows: ； in, is the positive definite inertia matrix; To use Simplify and organize The resulting column vector; express Relative to time The second derivative of .

4. The method for decentralized fault-tolerant control of a flexible joint space robot without velocity feedback according to claim 2, characterized in that: The specific implementation process of S3 is as follows: For the fast-changing subsystem in S2, in order to suppress the elastic vibration caused by the flexible joint, the torque differential feedback controller is designed as: ； in, is the diagonal coefficient matrix; is the defined singular perturbation factor; is the column vector of the control torque transmitted by the flexible joint, express Relative to time The first derivative of .

Images