Outer limb robot auxiliary support control method based on non-zero sum game

Abstract

The invention discloses an outer limb robot auxiliary support control method based on a non-zero sum game. The method comprises the following steps: according to the dynamics of a support control system, setting a state transition matrix of the system, inputting parameters of the matrix, initializing basic parameters, updating a state variable of the support control system and estimating deviation of the state variable; according to the Lyapunov stability control theorem and the state observer, an estimated value of the feedback gain of the wearer is obtained; calculating the optimal feedback gain of the outer limb according to the estimated values of the non-zero sum control, the game control and the wearer feedback gain; and according to the dynamic formula of the outer limb and the optimal feedback gain, the joint torque output of the outer limb robot is adjusted in a self-adaptive mode, and stable supporting of the outer limb robot is achieved. According to the method, the stability and smooth interaction of the outer limb robot during man-machine cooperation for supporting tasks can be improved, the success rate of overhead supporting operation is effectively increased, and the morbidity of musculoskeletal injuries of a user is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of external limb robots and human-machine coordinated control, and in particular to an auxiliary support control method of an external limb robot based on non-zero-sum game. Background Art

[0002] With the development of computer science, the degree of automation in industrial production is getting higher and higher, but due to factors such as process, cost, and space, many industrial production lines still have a large number of tasks that require manual work to complete. However, since such workers perform overhead work for a long time, they are prone to low work efficiency, muscle fatigue, and reduced physical fitness, and even occupational musculoskeletal injuries. Therefore, in order to improve this situation, using external limb robots to assist overhead tasks is an effective way to alleviate occupational musculoskeletal injuries.

[0003] An exo-limb robot is a robot worn on a human wearer that can perform tasks independently with an extra arm, especially in overhead support tasks, where the exo-limb robot can replace human hands for overhead support operations. However, since the exo-limb is worn on a person to replace the wearer's upper limbs to perform overhead tasks, this rigid connection makes the support stability of the exo-limb easily affected by the human wearer. In addition, since the output of the wearer and the exo-limb is difficult to coordinate, the unexpected force will react on the wearer's shoulder, which can easily cause further damage. Summary of the invention

[0004] The purpose of the present invention is to provide an external limb robot assisted support control method based on non-zero-sum game, which can effectively improve the stability of the external limb robot-assisted overhead support tasks, and can improve the flexibility of the equipment by adaptively adjusting the feedback gain of the external limb robot to reduce the reaction force on the wearer.

[0005] The present invention adopts the following technical solution:

[0006] An auxiliary support control method for an external limb robot based on non-zero-sum game, comprising the following steps:

[0007] S1. According to the dynamics of the supporting control system, set the parameters of the state transfer matrix A and input matrix B of the system, and initialize the basic parameters, which include state variables and state variable estimation deviations.

[0008] S2. Update the state variables and state variable estimation deviations supporting the control system.

[0009] S3. According to Lyapunov's stability control theorem and the state observer, the estimated value of the wearer's feedback gain is obtained.

[0010] S4. Calculate the optimal feedback gain of the external limb based on the estimated values ​​of non-zero-sum control, game control and wearer feedback gain.

[0011] S5. Based on the dynamics formula of the external limb and the optimal feedback gain, the joint torque output of the external limb robot is adaptively adjusted to achieve stable support of the external limb robot.

[0012] Furthermore, in step S1, the basic parameters also include the height h of the base of the external limb robot b , the height of the ceiling, the estimated value of the system's state variables, the mass m of the base of the external limb b , the mass of the support plate m p , mass of external limbs m s , the cost function parameter Q of the outer limb sl , the wearer's cost function parameter estimate Estimated value of wearer's feedback gain parameter The observation matrix parameter γ, the wearer feedback gain update parameter α and the lengths l1, l2, l3 of the link center of mass of the external limb to the joint.

[0013] Furthermore, in step S2, the update formula of the state variable and the state variable estimation deviation is:

[0014]

[0015]

[0016] Among them, h represents the height of the current support, Indicates the current speed of the support, h d represents the expected support height, represents the estimated value of the system's state variables, represents the estimated deviation of the updated state variable, and δ represents the updated state variable.

[0017] Furthermore, in step S3, updating the estimated value of the wearer feedback gain includes the following contents:

[0018] S301. The expression of the state observer supporting the control system is:

[0019]

[0020] in, represents the derivative of the estimated deviation of the state variable, P hl represents the wearer's feedback gain parameter, B T represents the transpose of B.

[0021] S302. The expression of Lyapunov function is:

[0022]

[0023] Where V represents the Lyapunov function, δ T represents the transpose of δ, express The transpose of , tr() represents the rank of the matrix, represents the deviation of the estimated value of the wearer's feedback gain, express The transpose of .

[0024] S303: Introducing the wearer input gain parameter update rate The specific expression is:

[0025]

[0026] S304: Update the wearer input gain parameter, the specific calculation method is:

[0027]

[0028] in, represents the wearer input gain parameter at time t, represents the wearer input gain parameter at time t-1, and Δt represents the time interval.

[0029] S305: Calculate the estimated value of the wearer's feedback gain. The specific formula is:

[0030]

[0031] in, Represents the estimated value of the wearer's feedback gain at time t.

[0032] Furthermore, in step S4, calculating the optimal feedback gain of the external limb includes the following contents:

[0033] When the wearer and the external limb coordinate to perform the overhead support task, the cost functions are:

[0034]

[0035]

[0036] in, represents the wearer's cost function, t0 represents the initial moment, represents the force exerted by the wearer on the movement of the support system, express The transpose of t represents the tth moment, J sl represents the cost function of the outer limb, u slrepresents the optimal output of the outer limb, Indicates u sl The transpose of .

[0037] The optimal control input of the exo-limb robot in the sense of Nash equilibrium is expressed as:

[0038] u sl =-Lδ

[0039] L=B T P sl

[0040] Where L represents the optimal feedback gain of the external limb; P sl Represents the feedback gain parameter of the robot.

[0041] P sl is the solution of the Riccati equation, and the specific expression is:

[0042]

[0043] Among them, A S represents the improved robot state transfer matrix, represents the estimated value of the wearer's feedback gain; Indicates A S The transpose of .

[0044] Furthermore, in step S5, the specific formula for adjusting the joint torque output of the external limb robot is:

[0045]

[0046] Wherein, τ represents the joint torque output of the exo-limb robot; M() represents the stargazing matrix; θ represents the joint position of the exo-limb robot; represents the acceleration of the external limb robot joints; represents the joint velocity of the external limb robot; C() represents the vector containing the Coriolis and centripetal moments; g() represents the vector containing the gravity moment; J represents the Jacobian matrix; J T represents the transpose of J; f represents the output force of the external limb robot, m p represents the mass of the support and g represents the acceleration due to gravity.

[0047] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0048] 1. The method proposed in the present invention estimates the wearer's input through an observer based on the system state and the Lyapunov stability control method, which not only reduces the system's dependence on force sensors, but also improves the stability of the control system.

[0049] 2. The method proposed in the present invention constructs the input force adjustment relationship between the wearer and the external limb robot through a non-cooperative game strategy, and can adjust the input force of the external limb robot according to the wearer's input feedback gain estimation, thereby improving the robot's coordination ability and compliance.

[0050] 3. The method proposed in the present invention is applicable to the adaptive adjustment strategy of the auxiliary overhead support control task of the exo-limb robot, which can adjust the support position of the exo-limb robot according to the wearer's movements and improve the robot's stable support ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is an overall implementation flow chart of the present invention.

[0052] Figure 2 It is a scene graph of an embodiment of the present invention. DETAILED DESCRIPTION

[0053] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0054] To achieve the above object, the present invention proposes a non-zero-sum game-based external limb robot auxiliary support control method, such as Figure 1 As shown, the specific steps are as follows:

[0055] S1. According to the dynamics of the support control system, set the parameters of the state transfer matrix A and input matrix B of the system, and initialize the basic parameters, which include state variables, state variable estimation deviation, and the height h of the base of the external limb robot. b , the height of the ceiling, the estimated value of the system's state variables, the mass m of the base of the external limb b , the mass of the support plate m p , mass of external limbs m s , the cost function parameter Q of the outer limb sl , the wearer's cost function parameter estimate Estimated value of wearer's feedback gain parameter The observation matrix parameter γ, the wearer feedback gain update parameter α and the lengths l1, l2, l3 of the link center of mass of the external limb to the joint.

[0056] S2, update the state variables and state variable estimation deviations of the supporting control system. The specific formula is:

[0057]

[0058]

[0059] Among them, h represents the height of the current support, Indicates the current speed of the support, h d represents the expected support height, represents the estimated value of the system's state variables, represents the estimated deviation of the updated state variable, and δ represents the updated state variable.

[0060] S3. According to Lyapunov's stability control theorem and the state observer, the estimated value of the wearer's feedback gain is obtained; the specific content is:

[0061] S301. The expression of the state observer supporting the control system is:

[0062]

[0063] in, represents the derivative of the estimated deviation of the state variable, P hl represents the wearer's feedback gain parameter, B T represents the transpose of B.

[0064] S302. The expression of Lyapunov function is:

[0065]

[0066] Where V represents the Lyapunov function, δ T represents the transpose of δ, surface

[0067] represents the transpose of δ, tr() represents the rank of the matrix, represents the deviation of the estimated value of the wearer's feedback gain, express The transpose of .

[0068] S303: In order to stabilize the system, the wearer inputs the gain parameter update rate. The specific expression is:

[0069]

[0070] S304: Update the wearer input gain parameter, the specific calculation method is:

[0071]

[0072] in, represents the wearer input gain parameter at time t, represents the wearer input gain parameter at time t-1, and Δt represents the time interval.

[0073] S305: Calculate the estimated value of the wearer's feedback gain. The specific formula is:

[0074]

[0075] in, Represents the estimated value of the wearer's feedback gain at time t.

[0076] S4. Calculate the optimal feedback gain of the external limb based on the estimated value of the non-zero-sum control, game control and the wearer's feedback gain; the specific content is:

[0077] When the wearer and the external limb coordinate to perform the overhead support task, the cost functions are:

[0078]

[0079]

[0080] in, represents the wearer's cost function, t0 represents the initial moment, represents the force exerted by the wearer on the movement of the support system, express The transpose of J sl represents the cost function of the outer limb, u sl represents the optimal output of the outer limb, Indicates u sl The transpose of .

[0081] The input of the cost function is optimal in the sense of Nash equilibrium, so the optimal control input of the exo-limb robot in the sense of Nash equilibrium is expressed as:

[0082] u sl =-Lδ

[0083] L=B T P sl

[0084] Where L represents the optimal feedback gain of the external limb; P sl Represents the feedback gain parameter of the robot.

[0085] P sl is the solution of the Riccati equation, and the specific expression is:

[0086]

[0087] Among them, A S represents the improved robot state transfer matrix, represents the estimated value of the wearer's feedback gain; Indicates AS The transpose of .

[0088] S5. Based on the dynamic formula of the external limb and the optimal feedback gain, the joint torque output of the external limb robot is adaptively adjusted to achieve stable support of the external limb robot; the specific contents are:

[0089]

[0090] Wherein, τ represents the joint torque output of the exo-limb robot; M() represents the stargazing matrix; θ represents the joint position of the exo-limb robot; represents the acceleration of the external limb robot joints; represents the joint velocity of the external limb robot; C() represents the vector containing the Coriolis and centripetal moments; g() represents the vector containing the gravity moment; J represents the Jacobian matrix; J T represents the transpose of J; f represents the output force of the external limb robot, m p represents the mass of the support and g represents the acceleration due to gravity.

[0091] Figure 2 (a) is a scene layout diagram of a specific implementation of the present invention, which includes a wearer, an external limb robot, an object to be supported (extruded board) and a ceiling. Figure 2 In (b), the wearer performs downward movement in the sagittal plane. At this time, the external limb robot performs auxiliary support control according to the detected human body movement position and speed through the proposed method. Figure 2 In (c), the wearer performs an upward movement in the sagittal plane, and the external limb is also controlled for auxiliary support according to the proposed method. According to the example demonstration, the method proposed in the present invention can enable the external limb robot to provide stable support and reduce the reaction force on the wearer.

[0092] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A non-zero-sum game-based external limb robot auxiliary support control method, characterized in that: include: S1. According to the dynamics of the supporting control system, set the parameters of the state transfer matrix A and the input matrix B of the system, and initialize the basic parameters, which include state variables and state variable estimation deviations; S2, updating the state variables and estimated deviations of the state variables supporting the control system; S3. According to Lyapunov stability control theorem and state observer, an estimated value of the wearer's feedback gain is obtained; S4, calculating the optimal feedback gain of the external limb according to the estimated values ​​of the non-zero-sum control, the game control and the wearer's feedback gain; S5. Based on the dynamics formula of the external limb and the optimal feedback gain, the joint torque output of the external limb robot is adaptively adjusted to achieve stable support of the external limb robot.

2. The non-zero-sum game-based external limb robot auxiliary support control method according to claim 1 is characterized in that: In step S1, the basic parameters also include the base height h of the external limb robot b , the height of the ceiling, the estimated value of the system's state variables, the mass m of the base of the external limb b , the mass of the support plate m p , mass of external limbs m s , the cost function parameter Q of the outer limb sl , the wearer's cost function parameter estimate Estimated value of wearer's feedback gain parameter The observation matrix parameter γ, the wearer feedback gain update parameter α and the lengths l1, l2, l3 of the link center of mass of the external limb to the joint.

3. The non-zero-sum game-based external limb robot auxiliary support control method according to claim 1, characterized in that: In step S2, the update formula of the state variable and the state variable estimation deviation is: Among them, h represents the height of the current support, Indicates the current speed of the support, h d represents the expected support height, represents the estimated value of the system's state variables, represents the estimated deviation of the updated state variable, and δ represents the updated state variable.

4. The non-zero-sum game-based external limb robot auxiliary support control method according to claim 1, characterized in that: In step S3, updating the estimated value of the wearer feedback gain includes the following: S301. The expression of the state observer supporting the control system is: in, represents the derivative of the estimated deviation of the state variable, P hl represents the wearer's feedback gain parameter, γ represents the measurement matrix parameter, represents the estimated deviation of the updated state variable, B T represents the transpose of B, represents the wearer's feedback gain parameter estimate, and δ represents the updated state variable; S302. The expression of Lyapunov function is: Where V represents the Lyapunov function, δ T represents the transpose of δ, express The transpose of , α represents the wearer feedback gain update parameter, tr() represents the rank of the matrix, represents the deviation of the estimated value of the wearer's feedback gain, express The transpose of S303: Introducing the wearer input gain parameter update rate The specific expression is: S304: Update the wearer input gain parameter, the specific calculation method is: in, represents the wearer input gain parameter at time t, represents the wearer input gain parameter at time t-1, Δt represents the time interval; S305: Calculate the estimated value of the wearer's feedback gain. The specific formula is: in, Represents the estimated value of the wearer's feedback gain at time t.

5. The non-zero-sum game-based external limb robot auxiliary support control method according to claim 1, characterized in that: In step S4, calculating the optimal feedback gain of the external limb includes the following contents: When the wearer and the external limb coordinate to perform the overhead support task, the cost functions are: in, represents the wearer's cost function, t0 represents the initial moment, represents the force exerted by the wearer on the movement of the support system, express The transpose of , t represents the tth moment, represents the wearer's cost function parameter estimate, δ represents the updated state variable, δ T represents the transpose of δ, J sl represents the cost function of the outer limb, Q sl represents the cost function parameter of the outer limb, u sl represents the optimal output of the outer limb, Indicates u sl The transpose of The optimal control input of the exo-limb robot in the sense of Nash equilibrium is expressed as: usl=-Lδ L=BTPs1 Where L represents the optimal feedback gain of the external limb; P sl Indicates the feedback gain parameter of the robot; B T represents the transpose of B; P sl is the solution of the Riccati equation, and the specific expression is: Among them, A S represents the improved robot state transfer matrix, represents the estimated value of the wearer's feedback gain; Indicates A S The transpose of .

6. The non-zero-sum game-based external limb robot auxiliary support control method according to claim 1, characterized in that: In step S5, the specific formula for adjusting the joint torque output of the external limb robot is: Wherein, τ represents the joint torque output of the exo-limb robot; M() represents the stargazing matrix; θ represents the joint position of the exo-limb robot; represents the acceleration of the external limb robot joints; represents the joint velocity of the external limb robot; C() represents the vector containing the Coriolis and centripetal moments; g() represents the vector containing the gravity moment; J represents the Jacobian matrix; J T represents the transpose of J; f represents the output force of the external limb robot, u sl represents the optimal output of the outer limb, m p represents the mass of the support and g represents the acceleration due to gravity.

Images