A hierarchical multi-loop control method for robotic arms for compliant assembly of electronic components

Abstract

The present invention provides a hierarchical multi-loop control method for a manipulator for compliant assembly of electronic components. The design reference trajectory, impedance force and the environmental impedance force of the previous cycle are input into the impedance feature knowledge base of the electronic component assembly scene to obtain the design impedance trajectory of the manipulator joint, and the difference is calculated with the position information of the manipulator joint. The control variable is predicted by the MPC cycle output model, and the auxiliary control variable is obtained by summing with the IHOSMC cycle output. The control torque of the manipulator system is obtained by the inverse dynamics control cycle equation, and then the joint angle variable and the output trajectory of the end effector are obtained. The environmental impedance force is obtained in combination with the environmental information; when the output trajectory of the end effector is equal to the design impedance trajectory and the environmental impedance force is equal to the design impedance force, the manipulator system is stable, and the assembly of electronic components is performed according to the output trajectory and environmental impedance force at this time. The present invention has the characteristics of high assembly accuracy and fast response speed.

Description

Technical Field

[0001] The invention relates to the field of robot arm automation technology and intelligent control, and in particular to a robot arm hierarchical multi-loop control method for flexible assembly of electronic components on printed circuit boards. Background Art

[0002] The robotic arm is a multi-input, multi-output, nonlinear, strongly coupled, time-varying dynamic system. During the service process, the joints have a series of uncertainties such as inaccurate structural parameters, friction, noise, interference, load changes, etc. The uncertainty of each joint will accumulate and expand at the end effector, which will seriously affect the service accuracy of the robotic arm.

[0003] However, the assembly scenarios of electronic components often require the assembly actions of the robot arm to be highly precise and have a fast response, which poses a challenge to the design of the robot arm control method. Currently, the most widely used control method in the industry is the PID-based control method. Since the pins of electronic components and the sockets of printed circuit boards are small and the pin types are complex and diverse, they are prone to collision and contact during the assembly process. In addition to requiring the assembly actions of the robot arm to have high precision and fast response, the assembly process is also required to have a certain degree of flexibility, that is, when the pins of the electronic components and the sockets of the printed circuit board are not axially aligned, the assembly action will not damage the pins, components or circuit boards. The PID control method is a control method based on error feedback, and it is difficult to achieve optimal control.

[0004] Therefore, in special service environments such as 1) electronic components are small, brittle and weak, 2) electronic components have large pin aspect ratios and are prone to plastic deformation, 3) printed circuit boards are highly integrated and easily damaged, and 4) for the sake of industry competitiveness, the shorter the assembly process cycle, the better; in order to ensure the efficiency, accuracy and safety of the electronic component assembly process, it is necessary to develop an optimal control method with higher accuracy and faster response for the online flexible assembly of electronic components by robotic arms. Summary of the invention

[0005] In view of this, the present invention provides a hierarchical multi-loop control method of a robot arm for flexible assembly of electronic components.

[0006] The present invention achieves the above technical objectives through the following technical means.

[0007] A hierarchical multi-loop control method for a robotic arm for compliant assembly of electronic components is provided, specifically:

[0008] The reference trajectory of the robot end effector, the design impedance force and the environmental impedance force of the previous cycle required for the electronic component assembly task are input into the electronic component assembly scene impedance feature knowledge base to obtain the design impedance trajectory of the robot end effector;

[0009] The difference between the designed impedance trajectory of the robot arm joint and the position information of the robot arm joint is passed to the model predictive control loop unit, and the model predictive control variable is output;

[0010] The model prediction control variable is summed with the integral high-order sliding mode control loop output signal to obtain an auxiliary control variable;

[0011] The auxiliary control variable is subjected to the inverse dynamics control cycle equation to obtain the control torque of the manipulator system, and then the joint angle variable is obtained. The output trajectory of the end effector is obtained from the joint angle variable, and then the environmental impedance is obtained in combination with the environmental information;

[0012] When the output trajectory of the end effector is equal to the designed impedance trajectory and the environmental impedance force is equal to the designed impedance force, the multi-joint robotic arm system is stable, and the robotic arm performs the assembly of electronic components according to the output trajectory and environmental impedance force at this time.

[0013] Furthermore, the electronic component assembly scenario impedance feature knowledge base includes experimental data of the process of a robotic arm assembling electronic components in different assembly methods and an adaptive neural fuzzy inference system trained using the experimental data, wherein the experimental data includes a reference trajectory, design impedance, environmental impedance and impedance trajectory.

[0014] Furthermore, the object of the model predictive control loop is the state space equation discretized after the integrated high-order sliding mode control loop And design the following cost function:

[0015]

[0016]

[0017] Where: The discrete form of matrix A Discrete form of matrix B T represents the sampling period of the discrete system, x i (k) represents the joint rotation information of the i-th joint at the k-th step of the discrete system, v i (k) represents the auxiliary control variable of the i-th joint at the k-th step of the discrete system; represents the 2-norm of a vector multiplied by the weight factor Q, Represents the 2-norm of a vector multiplied by the weight factor R, represents the 2-norm of a vector multiplied by the weight factor P; v i (k+j) is the control input of the i-th joint at the k-th control step, e i(k+j) is the error between the reference trajectory and the true trajectory of the i-th joint at the k-th control step, j represents the j-th prediction step, and N is the total prediction step length; x min and x max are the minimum and maximum values ​​set for the robot arm joint angles, v min and v max are the minimum driving torque value and the maximum driving torque value set for the joint angle of the robot arm respectively; i=1, 2, 3, n, where n is the total number of joints of the robot arm.

[0018] Furthermore, the model predicts that the control variable is in discrete form u MPC (k)=[v 1 (k),v 2 (k),…,v n (k)] T .

[0019] Furthermore, the integral high-order sliding mode control loop output signal is:

[0020]

[0021] The equivalent integral high-order sliding mode control output is:

[0022]

[0023] in: is greater than the auxiliary control variable v i (t) is the upper bound constant, sgn is the sign function, represents the integral sliding mode variable of the i-th joint, ξ i 2 (ζ) represents the derivative of the integrated sliding mode variable of the i-th joint.

[0024] Furthermore, the integral sliding mode variable and its derivative of the i-th joint satisfy:

[0025]

[0026] in: for The derivative of For i 2 The derivative of (t), represents the auxiliary control variable η of the i-th joint i (t) is the first-order derivative with respect to time, Represents the angular velocity of the i-th joint after decoupling The first derivative with respect to time, Represents the angular velocity of the i-th joint after decoupling The second derivative with respect to time, The integral term representing the integral sliding mode variable The second derivative with respect to time, m i is a positive real number, Represents the auxiliary control variable v of the i-th joint i (t) The first derivative with respect to time.

[0027] Furthermore, the auxiliary control variable v(t)=u MPC (t)+u IHOSMC (t), where u MPC (t) Continuous form of the model prediction control variables.

[0028] Furthermore, after the output trajectory of the end effector is subtracted from the environmental position information of the assembly scene, the difference is multiplied by the environmental stiffness, and the product obtained is the environmental impedance.

[0029] Furthermore, in the initial cycle, the environmental impedance is 0, the integral high-order sliding mode control loop output signal is 0, and the angular velocity of the joint is 0.

[0030] Furthermore, the designed impedance trajectory of the robot arm joint is obtained by the designed impedance trajectory of the robot arm end effector according to inverse kinematics.

[0031] The beneficial effects of the present invention are:

[0032] (1) The present invention obtains auxiliary control variables by solving model predictive control and integral high-order sliding mode control, obtains the control torque of the manipulator system through the inverse dynamics control loop equation, and then obtains the joint angle information, which greatly reduces the steady-state time of the manipulator during service and improves the steady-state accuracy and robustness of the system;

[0033] (2) The present invention decouples the dynamic equations of the manipulator system, decoupling a nonlinear coupling system into multiple equivalent linear single-input single-output decoupling systems, thereby greatly reducing the amount of online calculations required for the reasoning process while achieving optimal control;

[0034] (3) The over-integrated sliding mode control of the present invention eliminates uncertainties and interference factors, and can meet the needs of high-precision, fast-response online flexible assembly of electronic components in the presence of interference factors such as electromotive torque fluctuations, drive saturation, load changes, mechanical resonance, high-frequency disturbances, measurement errors and sampling delays. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a flow chart of the hierarchical multi-loop control method of a robotic arm for flexible assembly of electronic components according to the present invention;

[0036] Figure 2A schematic diagram is established for the impedance characteristic knowledge base of the electronic component assembly scenario of the present invention. DETAILED DESCRIPTION

[0037] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments, but the protection scope of the present invention is not limited thereto.

[0038] like Figure 1 As shown, the present invention provides a hierarchical multi-loop control method for a robot arm for flexible assembly of electronic components, which specifically includes the following steps:

[0039] Step (1): Establishment of the dynamic model of the general multi-joint robotic arm system

[0040] Obtain the inertia matrix M(q), centripetal force and Coriolis moment of the robot arm by consulting the robot arm structural parameters / technical data, measuring and calculating, or system identification. The gravity torque vector G(q) information is obtained, and the dynamic model of the general multi-joint robotic arm system is established as shown in the following formula:

[0041]

[0042] in: is an auxiliary variable, It is a comprehensive term composed of interference factors such as friction, disturbance, and uncertainty of load change. τ represents the motor torque of the robot arm, and q represents the joint angle. represents the joint angular velocity, Represents the joint angular acceleration.

[0043] Step (2), inverse dynamics control loop

[0044] To set the control rate based on inverse dynamics control:

[0045]

[0046] Where: v is the auxiliary control variable, and the symbol “^” represents the estimated value of the physical quantity it covers. Represents the joint angular velocity collected by the joint position sensor in real time.

[0047] Step (3), substitute equation (2) into equation (1) and rearrange it to obtain:

[0048]

[0049] in: It is an auxiliary intermediate variable. This physical quantity is generated due to the uncertainty factors of the multi-joint robot system.

[0050] Step (4), decoupling of the kinetic equations

[0051] Decouple equation (3) and let the decoupled joint angle and angular velocity vector We obtain multiple single-input single-output decoupling systems with the same number of robot joints as the number of robot arm joints, as shown below:

[0052]

[0053] in: and represents the rotation angle and angular velocity of the i-th joint after decoupling, and express and The first derivative with respect to time, v i (t) and η i (t) denotes the auxiliary control variable of the i-th joint.

[0054] Step (5), write equation (4) in the form of state space equation as follows:

[0055]

[0056] Among them: state quantity State quantity matrix matrix

[0057] Step (6): Design of hierarchical control strategy

[0058] The auxiliary control variable v(t) is decomposed into a combination of model predictive control and integral high-order sliding mode control in the following form:

[0059] v(t)=u MPC (t)+u IHOSMC (t) (6)

[0060] Step (7), integrated high-order sliding mode control (IHOSMC) loop

[0061] Assume the sliding manifold is:

[0062]

[0063] Where: m i is a positive real number, s i (t) is the sliding manifold of the i-th joint, e i (t) is the position error of the i-th joint, is the angular velocity error of the i-th joint, and n is the total number of joints of the robot.

[0064] Step (8), the integral sliding mode variable σ of the i-th joint i (t) can be designed as:

[0065]

[0066] in: is the integral term of the integral sliding mode variable, and s i (t 0 ) is t 0 The sliding manifold of the i-th joint at time, for From 0 The integral from time to time t.

[0067] Step (9), let have to:

[0068]

[0069] in: represents the integral sliding mode variable of the i-th joint, ξ i 2 (t) represents the derivative of the integrated sliding mode variable of the i-th joint, for The derivative of for The derivative of Represents η i (t) is the first-order derivative with respect to time, express The second derivative with respect to time, express The second derivative with respect to time, Represents the auxiliary control variable v of the i-th joint i (t) The first derivative with respect to time.

[0070] Step (10), the equivalent integral high-order sliding mode control output is:

[0071]

[0072] in: is greater than the control variable v i (t) is the upper bound constant, sgn represents the sign function.

[0073] Step (11), eliminate uncertainty and interference factors η through integral sliding mode control i (t), the state space equation (5) can be simplified to:

[0074]

[0075] Step (12), discretization of the state space equations

[0076] After the integrated high-order sliding mode control loop, the state space equation (11) can be discretized to obtain the following discrete form:

[0077]

[0078] in, T represents the sampling period of the discrete system, x i (k) represents the joint rotation information of the i-th joint at the k-th step of the discrete system, v i (k) represents the auxiliary control variable of the i-th joint at the k-th step of the discrete system, and the symbol “~” represents the discrete form of the physical quantity it covers.

[0079] Step (13), Model Predictive Control (MPC) loop

[0080] The control object of this loop is equation (12), let x ri is the reference trajectory of the ith joint, let e i =x ri -x i , then the cost function of model predictive control can be designed as follows:

[0081]

[0082] Among them: Symbol represents the 2-norm of a vector multiplied by the weight factor Q, represents the 2-norm of a vector multiplied by the weight factor R, Represents the 2-norm of a vector multiplied by the weight factor P, v i (k+j) is the control input of the i-th joint at the k-th control step, e i (k+j) is the error between the reference trajectory and the actual trajectory of the i-th joint at the k-th control step, j represents the j-th prediction step, and the total prediction step length is N; x min and x max are the minimum and maximum values ​​set for the robot arm joint angles, v min and v max They are the minimum and maximum driving torque values ​​set for the joint angles of the robot arm.

[0083] Step (14), installing the force sensor and the displacement sensor at the end effector of the robot arm to prepare for the establishment of the impedance characteristic knowledge base of the subsequent electronic component assembly scenario.

[0084] Step (15): Establishment of impedance characteristic knowledge base for electronic component assembly scenarios

[0085] like Figure 2 As shown, in the reference trajectory X r, Design impedance F d In a given scenario with assembly methods (such as clamping, pressing, rotating, pushing, embedding, etc.), the robot arm is dragged by the teaching method to complete the assembly process of electronic components, and the assembly method and reference trajectory X in the process are recorded. r , Design impedance F d , impedance locus X d and environmental impedance F e ; Among them, the environmental impedance F e The impedance trace X is recorded by the force sensor. d Then use the displacement sensor to record; after that, repeat the above experimental process a lot and make records to form a data set. The data set is divided according to different assembly methods, and the reference trajectory X in different assembly methods is r , Design impedance F d 、Environmental impedance F e and impedance locus X d The data is sorted and trained using the adaptive neural fuzzy inference system (ANFIS); the input parameter for ANFIS training is the reference trajectory X r , Design impedance F d and environmental impedance F e The output parameter for ANFIS training is the impedance trajectory X d The recorded experimental data and the trained network structure (ANFIS) are saved to establish the impedance characteristic knowledge base of the electronic component assembly scenario for subsequent use.

[0086] Step (16), such as Figure 1 As shown in the figure, the hierarchical multi-loop control method of the robot arm for the compliant assembly of electronic components is as follows:

[0087] In the initial cycle, the reference trajectory X of the robot end effector required for the electronic component assembly task is first designed. r (Where X r =[x r1 ,x r2 ,…,x rn ] T ) and design impedance F d , and then combined with the established electronic component assembly scene impedance feature knowledge base, the environmental impedance F under free conditions is obtained e =0) Design impedance trajectory X of the end effector of the robot arm d Then, according to inverse kinematics, the designed impedance trajectory q of the robot joint is d (t) is passed to the difference finder, which calculates the designed impedance trajectory q of the robot joint. d (t) and the initial position information q(t0 ) signal (collected by the displacement sensor on the joint) is subtracted and passed to the analog-to-digital converter, which then passes the discrete digital signal to the model predictive control (MPC) loop unit (the unit consists of n decoupled MPC controllers). The output of the loop is u MPC (k), where u MPC (k)=[v 1 (k),v 2 (k),…,v n (k)] T , after passing through the zero-order holder (ZOH), it is converted to u MPC (t); u MPC (t) and the integral high-order sliding mode control (IHOSMC) loop output signal is u IHOSMC (t)(and ) to sum (where u IHOSMC (t 0 )=0), and after passing through the summer, the auxiliary control variable v(t)=u in equation (6) is obtained MPC (t)+u IHOSMC (t), the control torque τ of the manipulator system is obtained through the inverse dynamics control cycle equation (2) (after the integral high-order sliding mode control (IHOSMC) cycle, the comprehensive term composed of interference factors such as uncertainty is eliminated; and in the initial cycle, ), τ is input into the robot arm to obtain the joint angle variable, which is converted into the spatial coordinate X information of the end effector (i.e., the output trajectory of the end effector) through forward kinematics, and then compared with the environmental position information X of the assembly scene. e Calculate the difference and compare the difference with the environmental stiffness K e Multiply them together, and the product is the environmental impedance F e (i.e. F e =K e ×(X e -X)), and then the design impedance F d Environmental resistance F e The data is input into the impedance characteristic knowledge base of the electronic component assembly scenario to close the initial cycle.

[0088] The designed impedance trajectory X of the end effector of the robot arm d According to inverse kinematics, the designed impedance trajectory q of the robot joint is obtained d The process of (t) is a prior art; the process of converting the joint angle variable information into the spatial coordinate X information of the end effector through forward kinematics is also a prior art.

[0089] The process of the subsequent cycle is similar to that of the initial cycle, but in the subsequent cycle, the electronic component assembly scenario impedance feature knowledge base input is: the reference trajectory X of the robot end effector required for the electronic component assembly task r , Design impedance F d The design impedance trajectory of the end effector of the manipulator under impedance conditions is obtained by combining the environmental impedance information obtained in the previous cycle, and the design impedance trajectory of the end effector of the manipulator under impedance conditions is subtracted from the real-time position information after inverse kinematics; and the signal output by the integral high-order sliding mode control (IHOSMC) loop is the signal of the previous cycle, Collected in real time by joint position sensors.

[0090] When the output trajectory of the robot arm is equal to its designed impedance trajectory X d , and the environmental impedance is equal to the design impedance F d When the multi-joint robotic arm system is stable, the robotic arm performs the assembly of electronic components according to the output trajectory and environmental impedance at this time.

[0091] The embodiments are preferred implementations of the present invention, but the present invention is not limited to the above-mentioned implementations. Any obvious improvements, substitutions or modifications that can be made by those skilled in the art without departing from the essential content of the present invention belong to the protection scope of the present invention.

Claims

1. A hierarchical multi-loop control method for a robotic arm for compliant assembly of electronic components, characterized in that: The reference trajectory of the robot end effector, the design impedance force and the environmental impedance force of the previous cycle required for the electronic component assembly task are input into the electronic component assembly scene impedance feature knowledge base to obtain the design impedance trajectory of the robot end effector; The difference between the designed impedance trajectory of the robot arm joint and the position information of the robot arm joint is passed to the model predictive control loop unit, and the model predictive control variable is output; The model prediction control variable is summed with the high-order sliding mode control loop output signal to obtain an auxiliary control variable; The auxiliary control variable is subjected to the inverse dynamics control cycle equation to obtain the control torque of the robot system, and then the joint angle variable is obtained. The output trajectory of the end effector is obtained from the joint angle variable, and then the environmental impedance is obtained in combination with the environmental information; When the output trajectory of the end effector is equal to the designed impedance trajectory and the environmental impedance force is equal to the designed impedance force, the multi-joint robotic arm system is stable, and the robotic arm performs the assembly of electronic components according to the output trajectory and environmental impedance force at this time.

2. The hierarchical multi-loop control method for a robotic arm according to claim 1, characterized in that: The electronic component assembly scenario impedance feature knowledge base includes experimental data of the process of a robotic arm assembling electronic components in different assembly methods and an adaptive neural fuzzy inference system trained using the experimental data, wherein the experimental data includes a reference trajectory, a design impedance, an environmental impedance, and an impedance trajectory.

3. The hierarchical multi-loop control method for a robotic arm according to claim 1, characterized in that: The object of the model predictive control loop is the state space equation discretized after the integrated high-order sliding mode control loop. And design the following cost function: Where: The discrete form of matrix A Discrete form of matrix B T represents the sampling period of the discrete system, x i (k) represents the joint rotation information of the i-th joint at the k-th step of the discrete system, v i (k) represents the auxiliary control variable of the i-th joint at the k-th step of the discrete system; represents the 2-norm of a vector multiplied by the weight factor Q, Represents the 2-norm of a vector multiplied by the weight factor R, represents the 2-norm of a vector multiplied by the weight factor P; v i (k+j) is the control input of the i-th joint at the k-th control step, e i (k+j) is the error between the reference trajectory and the true trajectory of the i-th joint at the k-th control step, j represents the j-th prediction step, and N is the total prediction step length; x min and x max are the minimum and maximum values ​​set for the robot arm joint angles, v min and v max are the minimum driving torque value and the maximum driving torque value set for the joint angle of the robot arm respectively; i=1, 2, 3, n, where n is the total number of joints of the robot arm.

4. The robot arm hierarchical multi-loop control method according to claim 3, characterized in that: The model predicts the control variable in discrete form as u MPC (k)=[v1(k),v2(k),…,v n (k)] T .

5. The robot arm hierarchical multi-loop control method according to claim 4, characterized in that: The output signal of the high-order sliding mode control loop is: The equivalent integral high-order sliding mode control output is: in: is greater than the auxiliary control variable v i (t) is the upper bound constant, sgn is the sign function, represents the integral sliding mode variable of the i-th joint, represents the derivative of the integrated sliding mode variable at the i-th joint.

6. The robot arm hierarchical multi-loop control method according to claim 5, characterized in that: The integral sliding mode variable and its derivative of the i-th joint satisfy: in: for The derivative of for The derivative of represents the auxiliary control variable η of the i-th joint i (t) is the first-order derivative with respect to time, Represents the angular velocity of the i-th joint after decoupling The first derivative with respect to time, Represents the angular velocity of the i-th joint after decoupling The second derivative with respect to time, The integral term representing the integral sliding mode variable The second derivative with respect to time, m i is a positive real number, Represents the auxiliary control variable v of the i-th joint i (t) The first derivative with respect to time.

7. The robot arm hierarchical multi-loop control method according to claim 5, characterized in that: The auxiliary control variable v(t)=u MPC (t)+u IHOSMC (t), where u MPC (t) Continuous form of the model prediction control variables.

8. The hierarchical multi-loop control method for a robotic arm according to claim 1, characterized in that: The output trajectory of the end effector is subtracted from the environmental position information of the assembly scene, and then multiplied by the environmental stiffness, and the product obtained is the environmental impedance.

9. The robot arm hierarchical multi-loop control method according to claim 1, characterized in that: In the initial cycle, the environmental impedance is 0, the high-order sliding mode control loop output signal is 0, and the angular velocity of the joint is 0.

10. The robot arm hierarchical multi-loop control method according to claim 1, characterized in that: The designed impedance trajectory of the robot arm joint is obtained by the designed impedance trajectory of the robot arm end effector according to inverse kinematics.

Images