Table tennis playing robot preset performance control method based on optimal human decision

Abstract

The invention relates to the technical field of robot intelligent control, and discloses a table tennis playing robot preset performance control method based on an optimal human decision, which comprises the following steps: reconstructing a kinetic model of a table tennis playing robot system, and obtaining servo constraints after state transformation; constructing a hierarchical hybrid controller of a robot preemption algorithm and a human decision control algorithm, integrating a human fuzzy intention with mechanical control to optimize a ball hitting strategy, and converting a decision problem into a functional optimization problem; and obtaining an analytical expression of the optimal membership function to obtain weights or priorities of human decisions in different states. According to the method, machine motion guided by a clear physical law is combined with fuzzy representation of human behaviors and intentions to form a coherent framework, optimization of system performance is achieved by integrating a preemption algorithm and a human decision algorithm, and meanwhile safety and robustness are ensured.

Description

Technical Field

[0001] This invention relates to the field of robot intelligent control technology, and in particular to a method for pre-setting performance control of a table tennis robot based on optimal human decision-making. Background Technology

[0002] The ping-pong robot is an intelligent device that integrates machine vision, motion control, and AI. By capturing the ball's trajectory and making real-time predictions, it completes precise shots through a mechanical structure. It can meet training and entertainment needs and is a typical example of the integration of technology and sports.

[0003] Existing robot control methods are mostly based on precise mathematical modeling and deterministic decision rules, relying on quantifiable and predictable input information to generate control commands. Although the intelligent decision-making algorithms in current robot control methods can analyze and recognize human intentions to a large extent through the analysis of electrophysiological signals such as electroencephalography (EEG), electrooculography (EOG), and electromyography (EMG), they struggle to capture and interpret the fuzziness and dynamic adaptability in human decision-making. For example, an athlete may adjust their hitting strategy in a very short time based on the opponent's hitting trajectory, ball speed, and spin. This decision-making process has a high degree of uncertainty and adaptability, which leads to existing table tennis robots often failing to accurately match this fuzzy decision-making logic, resulting in problems such as delayed response and a single adversarial strategy, making it difficult to achieve natural and efficient human-robot interaction. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides the following technical solution:

[0005] A method for pre-setting performance control of a table tennis robot based on optimal human decision-making includes the following steps:

[0006] S10: The constrained state is mapped to unconstrained state variables by the differential homeomorphism transformation method to reconstruct the dynamic model of the ping-pong robot system and obtain the servo constraints after the state transformation.

[0007] S20: Based on the UK method, a hierarchical hybrid controller consisting of a robot preemption algorithm and a human decision-making control algorithm is constructed, and a decision coordination module is introduced to resolve conflicts;

[0008] S30: Construct a Lagrange density function and integrate human fuzzy intentions with mechanical control by minimizing the action integral to optimize the hitting strategy;

[0009] S40: Transform the intelligent decision-making problem in the table tennis robot system into a functional optimization problem, and combine variational methods to obtain the analytical expression of the optimal membership function to obtain the weight or priority of human decisions under different states.

[0010] As an improvement to the above technical solution, the following steps need to be performed before step S10:

[0011] S01: Establish a dynamic model and performance constraints for a table tennis robot system that includes uncertain parameters and unknown environmental disturbances.

[0012] As an improvement to the above technical solution, step S10 specifically includes the following steps:

[0013] S11: Transform performance constraints into unconstrained state variables through a nonlinear transformation function;

[0014] S12: The zero-order constraints are transformed into first-order and second-order forms using a mapping transformation, and the dynamic model of the ping-pong robot system in the new coordinate system and the servo constraints after the state change are obtained.

[0015] As an improvement to the above technical solution, the nonlinear change function in step S11 includes the following formula:

[0016]

[0017] in, , , , and For the first The three transformation parameters corresponding to each joint are all constants.

[0018] As an improvement to the above technical solution, step S20 specifically includes the following steps:

[0019] S21: Construct performance constraint assumptions for the hierarchical hybrid control model and evaluate the constraint error of the system's tracking performance;

[0020] S22: Construct a hierarchical hybrid controller that includes machine preemptive algorithm control and human decision-making algorithm control based on performance constraint assumptions and constraint errors.

[0021] As an improvement to the above technical solution, the hierarchical hybrid controller includes the following formula:

[0022]

[0023] in, For policy algorithms, Indicates time, Represents coordinates, Let represent velocity, and be a scalar design parameter. >0.

[0024] As an improvement to the above technical solution, the Lagrange density function constructed in step S30 includes the following formula:

[0025]

[0026] in, It is an integration of vague human intentions and mechanical control. For mechanical systems, For fuzzy terms related to human intent;

[0027] The action integral minimized in step S30 is as follows:

[0028]

[0029] As an improvement to the above technical solution, step S40 specifically includes the following steps:

[0030] Step S41: Construct the functional and its domain and boundary conditions;

[0031] Step S42: Construct a membership function, and construct the cost functional based on the implementation conditions of the membership function and the functional.

[0032] Step S43: Based on the symmetry of the domain at the mass point, obtain the analytical expression of the optimal membership function within the domain.

[0033] As an improvement to the above technical solution, the implementation conditions of the membership function are as follows:

[0034] The membership function is continuous and symmetric with respect to the centroid of the domain.

[0035] As an improvement to the above technical solution, the performance of the hierarchical hybrid controller is verified using the following steps:

[0036] S50: Numerical simulation experiments were conducted on the output of the hierarchical hybrid controller using MATLAB to verify the robust control performance based on UK theory.

[0037] The beneficial effects of this invention are:

[0038] This pre-set performance control method for a table tennis robot based on optimal human decision-making combines machine motion guided by explicit physical laws with fuzzy representations of human behavior and intentions to form a coherent framework. By integrating preemption algorithms and human decision-making algorithms, it achieves optimal system performance while ensuring safety and robustness. Attached Figure Description

[0039] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0040] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention.

[0041] A method for pre-setting performance control of a table tennis robot based on optimal human decision-making includes the following steps:

[0042] Step S01: Establish a dynamic model and performance constraints for the table tennis robot system, including uncertain parameters and unknown environmental disturbances.

[0043] Consider a ping-pong robot system. The system's control task is to combine mechanical motion with human fuzzy perception to control the end effector to track the trajectory of the target ping-pong ball. Simplifying the system's dynamics analysis, we can obtain the dynamic model of the ping-pong robot system, as follows:

[0044]

[0045] in, Represents coordinates, Indicates speed, Indicates acceleration. Represents the inertia matrix. Represents the Coriolis and centrifugal force matrices. Represents gravity. Indicates the input matrix, and furthermore, It is an uncertain parameter. It indicates disturbances in an unknown environment. The set represents the control input (whose magnitude may depend on the human decision-making control area). and Both are compact sets, representing respectively and unknown disturbances The possible regions where the function is located , , and All are continuous.

[0046] The dynamic model of the table tennis robot system is subject to the following servo constraints:

[0047]

[0048] in, It is the right-hand side of the constraint equation. These are the coefficients of the constraint equations.

[0049] and All are continuously differentiable of one order, and This constraint can be written in matrix form:

[0050]

[0051] in, ,

[0052] ;

[0053] Taking the derivative with respect to t, we obtain the constraint conditions:

[0054]

[0055] in, Indicates the original position constraint For generalized coordinates The Jacobian matrix obtained by taking the partial derivatives. Represents the original position-level constraint function The total differential along the feasible motion, yes The One portion, and All The above constraints can be expressed in matrix form:

[0056]

[0057] in, and .

[0058] Use the task trajectory of the ping-pong robot as a constraint The mathematical definition is as follows:

[0059]

[0060] Step S10: The constrained state is mapped to an unconstrained state variable by using the differential homeomorphism transformation method to reconstruct the dynamic model of the ping-pong robot system and obtain the servo constraints after the state transformation.

[0061] Step S10 specifically includes the following steps:

[0062] Step S11: Transform performance constraints into unconstrained state variables using a nonlinear transformation function;

[0063] State variables in the dynamic model of a table tennis robot system Subject to bilateral inequalities, it can be expressed as:

[0064] ,

[0065] in, and It is a constant;

[0066] Regarding the above state variables The bilateral inequality constraints are addressed by using the differential homeomorphism method to transform the constrained state variables. Transform into unconstrained state variables .

[0067] Consider the following nonlinear variation function:

[0068]

[0069] in, , and For the first The three transformation parameters corresponding to each joint are all constants, and , .

[0070] system , and It can be calculated using the following three conditions:

[0071]

[0072]

[0073]

[0074] From the above three equations, we can derive

[0075]

[0076]

[0077]

[0078] Step S12: Use mapping transformation to convert the zero-order constraints into first-order and second-order forms, and obtain the dynamic model of the ping-pong robot system in the new coordinate system and the servo constraints after the state change.

[0079] By using mapping transformation, Transform into and :

[0080]

[0081]

[0082]

[0083] in, , Used to implement the original physical joint angle variable Mapping to a new unconstrained control variable.

[0084] In this process, the control design of constrained dynamics can be transformed into the control design of unconstrained dynamics. For the sake of simplicity, all subsequent formulas refer to the first... Each joint is established, and the conversion is used. Therefore, we get:

[0085]

[0086]

[0087]

[0088] By transforming the above equation, the dynamic model of the ping-pong robot system in the new coordinate system can be explicitly expressed as:

[0089]

[0090] in, Represents coordinates, Indicates speed, Indicates acceleration. Represents the inertia matrix. Indicates the input matrix, and furthermore, It represents the sum of all conservative and non-conservative generalized forces, including gravity, joint friction, cable force, and external disturbances. It is an uncertain parameter. It indicates disturbances in an unknown environment. The set represents the control input (whose magnitude may depend on the human decision-making control area). and Both are compact sets, representing respectively and unknown disturbances The possible regions where the function is located ,and All are continuous.

[0091] The coefficient matrices and vectors are defined as follows:

[0092]

[0093]

[0094]

[0095] Equivalently, the servo constraints after the state transformation become:

[0096]

[0097]

[0098]

[0099] in, and This refers to the direct mapping of position-level servo constraints to the new coordinates. and This refers to the Jacobian mapping of the velocity-level servo constraint in the new coordinate system. This refers to the complete expression of the acceleration level servo constraint in the new coordinate system. Taking the first-order differential of the constraint equations, The constraint equations were differentiated into second order, and A, b, and c are the corresponding constraint matrices.

[0100] The quantities are defined as follows:

[0101]

[0102]

[0103]

[0104]

[0105] in, , , , and The wavy lines above represent the original functions or physical quantities before the coordinate transformation.

[0106] After the above transformation, the original finite field (bounded) inequality joint constraints are transformed into constraints on open intervals (i.e., the state variables are now located in an infinite field). With the help of the given differential homeomorphism transformation, the finite field state variables are mapped to the infinite field state variables, thus embedding them into the dynamic model of the new ping-pong robot system.

[0107] Step S20: Based on the UK method, construct a hierarchical hybrid controller consisting of a robot preemption algorithm and a human decision control algorithm.

[0108] Step S20 specifically includes the following steps:

[0109] Step S21: Construct the performance constraint assumptions of the hierarchical hybrid control model and evaluate the constraint error of the system tracking performance.

[0110] The following performance constraint assumptions are proposed:

[0111] Assumption 1: For each and They all ,and The "nominal" part of the inertia matrix satisfies And the function continuous.

[0112] definition:

[0113]

[0114]

[0115]

[0116] Combining the three equations, we get: ;

[0117] Assumption 2: For each , It is at full capacity.

[0118] Assumption 3: Based on Assumption 2, for a given , ,make Then for all , there are parameters , making .

[0119] in, A diagonal matrix with freely chosen positive constants. 0 is a global positive constant that is only related to the upper bound of uncertainty.

[0120] The core of the proposed method is to input the environmental disturbance into the matrix. Decompose the system to show the impact of environmental uncertainties on the system.

[0121] This decomposition is performed under the premise that assumption 2 holds, ensuring that Reversible.

[0122] Therefore,

[0123] in,

[0124]

[0125]

[0126] Assumption 4: There exists a constant vector It is known or unknown, and a known function. : This makes it possible for all

[0127]

[0128] make , is used to represent the constraint error used to evaluate the tracking performance of the system.

[0129] Step S22: Construct a hierarchical hybrid controller that includes machine preemption algorithm control and human decision-making algorithm control based on performance constraint assumptions and constraint errors.

[0130] Under assumption 2, let... .

[0131] Under assumption 4, let... .

[0132] The following robust control measures are proposed:

[0133]

[0134] in, These are scalar design parameters, and their specific values ​​can be determined based on human decision-making.

[0135] If assumptions 1 to 4 above are satisfied, the proposed control strategy can enable the system to achieve uniform boundedness and uniform final boundedness.

[0136] exist Part 1 + The strategy algorithm is a mature algorithm that a computer can implement. The choice of algorithm depends on the requirements of the control strategy. (Part Two) Then one can choose based on human judgment. The control design requirements in the selection This parameter must be greater than zero, but there are no other restrictions.

[0137] Step S30: Construct the Lagrange density function and integrate human fuzzy intentions with mechanical control by minimizing the action integral to optimize the hitting strategy;

[0138] Construct the Lagrange density function:

[0139] in, For mechanical systems, For fuzzy terms related to human intent.

[0140] By minimizing the action integral It integrates vague human intentions with mechanical control to optimize ball-striking strategies.

[0141] Step S40: Transform the intelligent decision-making problem in the ping-pong robot system into a functional optimization problem, and use variational methods to obtain the analytical expression of the optimal membership function to obtain the weight or priority of human decisions under different states.

[0142] Step S40 specifically includes the following steps:

[0143] Step S41: Construct the functional and its domain and boundary conditions.

[0144] Consider the following functional:

[0145]

[0146] in, It is a Lagrange quantity, and And the boundary conditions are satisfied: , .

[0147] Assumption (Differentiate twice), then any function Minimize functional And satisfy the boundary conditions , It should satisfy the following Euler-Lagrange equations:

[0148]

[0149] set up In practice, The realization of has limits, therefore, let and They are The upper and lower bounds, of which .

[0150] Step S42: Construct a membership function, and construct the cost functional based on the implementation conditions of the membership function and the functional.

[0151] Consider a membership function : .

[0152] The following conditions must be met:

[0153] 1. Function It is continuous;

[0154] 2. It is relative to the center of mass. Symmetry, specifically ,in ;

[0155] 3. When , ,therefore .

[0156] The above optimal membership function Subject to the upper realm and the lower realm Restrictions, among which , Satisfying continuity and relative to the center of mass Symmetry;

[0157] Considering ,in It is a differential operator. express Arc length on a plane.

[0158]

[0159] Combining the optimal membership function described above, consider the cost functional: It represents the curve length. .

[0160] Specifically, it is expressed as follows:

[0161]

[0162] Step S43: Based on the symmetry of the domain at the mass point, obtain the analytical expression of the optimal membership function within the domain.

[0163] Depend on We can obtain, , Applying the Euler-Lagrange equations, we obtain:

[0164]

[0165] Direct differentiation yields:

[0166]

[0167]

[0168]

[0169]

[0170]

[0171] The denominator is not zero, therefore we know ,Right now .

[0172] Depend on , We can obtain, .

[0173] Here , can be obtained

[0174] This can be expressed as:

[0175]

[0176] Using boundary conditions It can be restated as:

[0177]

[0178] exist At that point, another boundary condition is ,have:

[0179]

[0180] It can be solved The is:

[0181]

[0182] Finally, for all The membership function is:

[0183]

[0184] Final cost for:

[0185]

[0186] based on The symmetry of all The optimal membership function is as follows:

[0187] .

[0188] The performance of the hierarchical hybrid controller was verified using the following steps:

[0189] Step S50: Use MATLAB to conduct numerical simulation experiments on the output of the hierarchical hybrid controller to verify the robust control performance based on UK theory.

[0190] Numerical simulations were performed using MATLAB in three scenarios to verify the effectiveness, necessity, and applicability of the proposed algorithm. In these simulations, and Treated as an uncertain parameter (i.e.) , ,in , The parameter uncertainties and environmental interference settings for each scenario are kept consistent to ensure the fairness of the comparison.

[0191] The above embodiments are merely illustrative of the technical solutions of the present invention and are not intended to limit it. Anyone skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A preset performance control method for a table tennis robot based on optimal human decision-making, characterized in that, The method comprises the following steps: S10: mapping the restricted state into an unconstrained state variable by a differential homeomorphism transformation method to reconstruct the dynamic model of the table tennis robot system and obtain the servo constraint after state transformation; S20: constructing a hierarchical hybrid controller composed of the robot preemption algorithm and the human decision control algorithm based on the U-K method, and introducing a decision coordination module to resolve the conflict; S30: constructing a Lagrange density function and integrating the human fuzzy intention and the mechanical control by minimizing the action integral to optimize the ball hitting strategy; S40: converting the intelligent decision problem in the table tennis robot system into a functional optimization problem, and obtaining the analytical expression of the optimal membership function by combining the variational method to obtain the weight or priority of the human decision under different states.

2. The preset performance control method for a table tennis robot based on optimal human decision according to claim 1, characterized in that: Before the step S10 is performed, the following step is further performed: S01: establishing the dynamic model and performance constraint of the table tennis robot system containing uncertain parameters and unknown environmental disturbances for the table tennis robot system.

3. The preset performance control method for a table tennis robot based on optimal human decision according to claim 1, characterized in that: The step S10 specifically comprises the following steps: S11: converting the performance constraint into an unconstrained state variable by a nonlinear transformation function; S12: converting the zero-order constraint condition into the first-order and second-order forms by a mapping transformation method, and obtaining the dynamic model of the table tennis robot system in the new coordinate system and the servo constraint after state transformation.

4. The preset performance control method for a table tennis robot based on optimal human decision according to claim 3, characterized in that: The nonlinear transformation function in the step S11 comprises the following formula: in, , , , and For the first The three transformation parameters corresponding to each joint are all constants.

5. The preset performance control method for a table tennis robot based on optimal human decision according to claim 1, characterized in that: The step S20 specifically comprises the following steps: S21: constructing the performance constraint hypothesis of the hierarchical hybrid control model, and evaluating the constraint error of the system tracking performance; S22: constructing the hierarchical hybrid controller including the robot preemption algorithm control and the human decision algorithm control according to the performance constraint hypothesis and the constraint error.

6. The preset performance control method for a table tennis robot based on optimal human decision according to claim 5, characterized in that: The hierarchical hybrid controller comprises the following formula: wherein, is a policy algorithm, denotes time, denotes coordinates, denotes velocity, k is a scalar design parameter, and >

0.

7. The preset performance control method for a table tennis robot based on optimal human decision according to claim 1, characterized in that: The Lagrange density function constructed in the step S30 comprises the following formula: wherein, is a human fuzzy intent and mechanical control integration term, is a mechanical system term, is a human intent fuzziness term; The minimizing action integral in the step S30 is as follows: 。 8. The preset performance control method for a table tennis robot based on optimal human decision according to claim 1, characterized in that: The step S40 specifically comprises the following steps: Step S41: constructing the functional and the definition domain and boundary condition of the functional; Step S42: constructing the membership function, and constructing the cost functional according to the implementation condition of the membership function and the functional; Step S43: obtaining the analytical expression of the optimal membership function in the definition domain according to the symmetry of the definition domain of the mass point.

9. The pre-set performance control method for a table tennis robot based on optimal human decision according to claim 8, characterized in that: The implementation condition of the membership function is as follows: The membership function is continuous, and the membership function is symmetric with respect to the centroid of the definition domain.

10. The pre-set performance control method for table tennis playing robots based on optimal human decision according to any one of claims 1-9, characterized in that, The performance of the hierarchical hybrid controller is verified by the following steps: S50: performing a numerical simulation experiment on the output of the hierarchical hybrid controller by MATLAB to verify the robust control performance based on the UK theory.

Images