阅读说明：本技术 Scara机器人的摩擦模型改进以及动力学参数辨识方法 (The friction model of SCARA robot improves and dynamic parameters identification method ) 是由 袁野 白瑞林 李新 于 2019-09-16 设计创作，主要内容包括：本发明公开了一种SCARA机器人的摩擦模型改进以及动力学参数辨识方法,包括：建立SCARA机器人关节动力学模型；对改进的摩擦模型进行参数辨识；将辨识后的所述摩擦模型代入动力学模型,对所述动力学模型中除所述摩擦模型外的剩余部分线性化；基于线性化后的所述动力学模型,设定观测矩阵和限制条件,从而设计出改进傅里叶形式的激励轨迹；基于所述激励轨迹,通过实验采集对应的数据,从而获得待辨识的动力学参数；通过最小二乘法辨识所述待辨识的动力学参数。实现提高参数辨识的精度和力矩预测的准确性的优点。(The invention discloses a kind of improvement of the friction model of SCARA robot and dynamic parameters identification methods, comprising: establishes SCARA joint of robot kinetic model；Parameter identification is carried out to improved friction model；The friction model after identification is substituted into kinetic model, the remainder in the kinetic model in addition to the friction model is linearized；Based on the kinetic model after linearisation, observing matrix and restrictive condition are set, to design the excitation track for improving Fourier formalism；Based on the excitation track, corresponding data are acquired by experiment, to obtain kinetic parameter to be identified；Pass through kinetic parameter to be identified described in least squares identification.The advantages of realizing the accuracy that the precision for improving parameter identification and torque are predicted.)

1. a kind of friction model of SCARA robot improves and dynamic parameters identification method characterized by comprising

Establish SCARA joint of robot kinetic model；

Parameter identification is carried out to improved friction model；

By after identification the friction model substitute into kinetic model, in the kinetic model in addition to the friction model Remainder linearisation；

Based on the kinetic model after linearisation, observing matrix and restrictive condition are set, improves Fourier to design The excitation track of form；

Based on the excitation track, corresponding data are acquired by experiment, to obtain kinetic parameter to be identified；

Pass through kinetic parameter to be identified described in least squares identification.

2. the friction model of SCARA robot according to claim 1 improves and dynamic parameters identification method, special Sign is, described to establish SCARA joint of robot kinetic model, comprising:

Simplified SCARA Dynamic Models of Robot Manipulators is established with Lagrangian method；

The kinetics equation in kinetic model that the SCARA robot in the joint n simplifies are as follows:

In formula, q is corner vector, and 1,2 ranks are ledRespectively angular speed and angular acceleration vector, H (q) be n rank machine Device people's inertial matrix,For coriolis force centrifugal force matrix, G (q) is gravity vector,τ is respectively to rub, drive Kinetic moment vector.

3. the friction model of SCARA robot according to claim 2 improves and dynamic parameters identification method, special Sign is, described to carry out parameter identification to improved friction model, comprising:

Experimental data is acquired under setting condition；

Axis moment of friction improved friction mould associated with axis angular rate, axle acceleration and Angle Position is established based on experimental data Type；

Parameter identification is carried out based on the improved friction model.

4. the friction model of SCARA robot according to claim 3 improves and dynamic parameters identification method, special Sign is, acquires experimental data under setting condition are as follows:

The constant speed tracking test of uniaxial 36 groups of friction speeds in the output revolving speed section of motor side is acquired under setting condition Data.

5. the friction model of SCARA robot according to claim 4 improves and dynamic parameters identification method, special Sign is, establishes the associated improved friction model of axis moment of friction and axis angular rate, axle acceleration, packet based on experimental data It includes:

Experimental data is fitted by coulomb+viscid friction model:

Wherein, F_vFor frictional force relevant to speed, f_cFor Coulomb friction coefficient, f_vFor viscous friction coefficient；

Friction model is decomposed into friction model relevant to speed and friction model relevant with angle；

Friction model relevant to the speed adds friction term and high-speed friction compensation on the basis of coulomb+viscid friction Are as follows:

Wherein f_sFor confficient of static friction, v_sFor speed proportional coefficient, f_aIt is the high-speed friction penalty coefficient of addition.

6. the friction model of SCARA robot according to claim 5 improves and dynamic parameters identification method, special Sign is, establishes axis moment of friction improved friction model associated with Angle Position based on experimental data, comprising:

The experimental data is subjected to fast Fourier variation when friction model relevant to the angle is analyzed, obtains dominant pilot Rate ingredient is indicated using SIN function combining form are as follows:

Wherein, F_pFor frictional force relevant to Angle Position, A₁, A₂For amplitude,For phase shift, p is the angle position for inputting motor side It moves.

7. the friction model of SCARA robot according to claim 6 improves and dynamic parameters identification method, special Sign is that the kinetic model based on after linearisation sets observing matrix and restrictive condition, to design improvement The excitation track of Fourier formalism, comprising:

Construct the optimisation criteria of observing matrix are as follows:

Wherein, Y is observing matrix, and the minimum singular value of σ min (Y) representing matrix, λ expression weight, λ is equal to 0.1.

8. the friction model of SCARA robot according to claim 7 improves and dynamic parameters identification method, special Sign is that the kinetic model based on after linearisation sets observing matrix and restrictive condition, to design improvement The excitation track of Fourier formalism, comprising:

The limitation of angle, velocity and acceleration in the excitation track is solved by genetic algorithm；

For the individual for violating constraint in genetic algorithm, apply a punishment letter on the fitness value for the individual for violating constraint Number, the penalty are as follows:

F '=F+ α max { 0, p },

F is former fitness function, and α is a penalty function factor greater than 0, and p is penalty function, and being unsatisfactory for p when constraint is positive value, full P is 0 when sufficient.

9. the friction model of SCARA robot according to claim 8 improves and dynamic parameters identification method, special Sign is, described to be based on the excitation track, acquires corresponding data by experiment, to obtain dynamics ginseng to be identified Number, comprising:

It is handled using joint angles of the Nonlinear Tracking Differentiator to acquisition, to seek angular velocity signal and angular acceleration Signal,

The Nonlinear Tracking Differentiator includes velocity factor r and high sp eed and optimal control comprehensive function u (x₁,x₂)。

10. the friction model of SCARA robot according to claim 9 improves and dynamic parameters identification method, It is characterized in that, it is described to pass through kinetic parameter to be identified described in least squares identification, comprising:

By least square method to the Chemical kinetic parameter estimation to be identified:

Wherein, Y is observing matrix,Respectively angular speed and angular acceleration vector, τ are driving moment vector,To need The kinetic parameter collection of identification, F_fFor friction model.

Technical field

The present invention relates to the dynamics Controllings of SCARA robot to optimize field, and in particular, to SCARA robot rubs Wipe model refinement and dynamic parameters identification method.

Background technique

With the continuous development of intelligence manufacture industry, industrial robot is gradually applied to high-precision field.Currently, domestic Industrial robot mostly uses greatly based on kinematic control method, and each joint uses independent PID control strategy, and tracking accuracy is not It is high.The Dynamic matrix control based on model is designed, considers that the dynamics in each joint is to realize the effective ways of high-precision motion control, But such control strategy requires the accurate kinetic model and kinetic parameter of robot.

The acquisition methods of robot dynamics' parameter mainly have disintegration mensuration, CAD method and whole identification method.Robot Structure is complicated, and disintegration mensuration can not directly measure all parameters；CAD method has ignored the equipment error of robot, precision It is not high；Whole identification method does not need dismantling robot, and without special experiment porch is built, process is convenient, receives extensively Using.But existing identification method there is a problem of parameter identification precision not enough and torque prediction accuracy deficiency.

Summary of the invention

It is an object of the present invention in view of the above-mentioned problems, propose that a kind of friction model of SCARA robot is improved and moved Mechanics parameter discrimination method, with realize improve parameter identification precision and torque prediction accuracy the advantages of.

To achieve the above object, technical solution used in the embodiment of the present invention is:

A kind of the friction model improvement and dynamic parameters identification method of SCARA robot, comprising:

Establish SCARA joint of robot kinetic model；

Parameter identification is carried out to improved friction model；

By after identification the friction model substitute into kinetic model, in the kinetic model remove the friction model Outer remainder linearisation；

Based on the kinetic model after linearisation, observing matrix and restrictive condition are set, improves Fu to design In leaf form excitation track；

Based on the excitation track, corresponding data are acquired by experiment, to obtain kinetic parameter to be identified；

Pass through kinetic parameter to be identified described in least squares identification.

It is further, described to establish SCARA joint of robot kinetic model, comprising:

Simplified SCARA Dynamic Models of Robot Manipulators is established with Lagrangian method；

The kinetics equation in kinetic model that the SCARA robot in the joint n simplifies are as follows:

In formula, q is corner vector, and 1,2 ranks are ledRespectively angular speed and angular acceleration vector, H (q) be n Rank robot inertial matrix,For coriolis force centrifugal force matrix, G (q) is gravity vector,τ is respectively to rub It wipes, driving moment vector.

It is further, described that parameter identification is carried out to improved friction model, comprising:

Experimental data is acquired under setting condition；

Establish that axis moment of friction is associated with axis angular rate, axle acceleration and Angle Position improved to rub based on experimental data Wipe model；

Parameter identification is carried out based on the improved friction model.

Further, experimental data is acquired under setting condition are as follows:

The constant speed tracking of uniaxial 36 groups of friction speeds in the output revolving speed section of motor side is acquired under setting condition Experimental data.

Further, axis moment of friction is established based on experimental data and axis angular rate, axle acceleration is associated improved Friction model, comprising:

Experimental data is fitted by coulomb+viscid friction model:

Wherein, F_vFor frictional force relevant to speed, f_cFor Coulomb friction coefficient, f_vFor viscous friction coefficient；

Friction model is decomposed into friction model relevant to speed and friction model relevant with angle；

Friction model relevant to the speed adds friction term and high-speed friction on the basis of coulomb+viscid friction Compensation term are as follows:

Wherein f_sFor confficient of static friction, v_sFor speed proportional coefficient, f_aIt is the high-speed friction penalty coefficient of addition.

Further, axis moment of friction improved friction model associated with Angle Position is established based on experimental data, wrapped It includes:

The experimental data is subjected to fast Fourier variation when friction model relevant to the angle is analyzed, is led Setting frequency ingredient is indicated using SIN function combining form are as follows:

Wherein, F_pFor frictional force relevant to Angle Position, A₁, A₂For amplitude,For phase shift, p is input motor side Angular displacement.

Further, the kinetic model based on after linearisation sets observing matrix and restrictive condition, thus Design the excitation track for improving Fourier formalism, comprising:

Construct the optimisation criteria of observing matrix are as follows:

Wherein, Y is observing matrix, and the minimum singular value of σ min (Y) representing matrix, λ expression weight, λ is equal to 0.1.

Further, the kinetic model based on after linearisation sets observing matrix and restrictive condition, thus Design the excitation track for improving Fourier formalism, comprising:

The limitation of angle, velocity and acceleration in the excitation track is solved by genetic algorithm；

For the individual for violating constraint in genetic algorithm, applies one on the fitness value for the individual for violating constraint and punish Penalty function, the penalty are as follows:

F '=F+ α max { 0, p },

F is former fitness function, and α is a penalty function factor greater than 0, and p is penalty function, and p is positive when being unsatisfactory for constraint Value, p is 0 when meeting.

Further, described to be based on the excitation track, corresponding data are acquired by experiment, to obtain to be identified Kinetic parameter, comprising:

It is handled using joint angles of the Nonlinear Tracking Differentiator to acquisition, is added to seek angular velocity signal with angle Speed signal,

The Nonlinear Tracking Differentiator includes velocity factor r and high sp eed and optimal control comprehensive function u (x₁,x₂)。

It is further, described to pass through kinetic parameter to be identified described in least squares identification, comprising:

By least square method to the Chemical kinetic parameter estimation to be identified:

Wherein, Y is observing matrix,Respectively angular speed and angular acceleration vector, τ are driving moment vector, To need the kinetic parameter collection recognized, F_fFor friction model.

Technical solution of the present invention has the advantages that

(1) improved friction model is added in robot body kinetic model to obtain completely improving kinetic simulation Type, this accuracy for predicting the precision for improving parameter identification and torque.

(2) the complex nonlinear phenomenon to rub when being directed to SCARA robot motion, improvement friction model proposed by the present invention Complicated friction phenomenon when SCARA robot high-speed motion can preferably be characterized.

(3) by combinatorial optimization algorithm of the design with Reward-Penalty Functions, excitation track used when parameter identification is devised, is added The fast optimization time, and obtain recognizing the signals such as required joint velocity acceleration and torque by using data prediction.This Invention improves the precision integrally recognized while guaranteeing that SCARA robot motion is not transfinited.

Below by drawings and examples, technical scheme of the present invention will be described in further detail.

Detailed description of the invention

Fig. 1 is friction model improvement and the dynamic parameters identification side of SCARA robot described in the embodiment of the present invention The flow chart of method；

Fig. 2 is SCARA robot links coordinate system diagram described in the embodiment of the present invention；

Fig. 3 is SCARA robot architecture's schematic diagram described in the embodiment of the present invention；

Fig. 4 is excitation track combination optimization algorithm flow chart described in the embodiment of the present invention.

Specific embodiment

Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings, it should be understood that preferred reality described herein Apply example only for the purpose of illustrating and explaining the present invention and is not intended to limit the present invention.

As shown in Figure 1, a kind of friction model of SCARA robot improves and dynamic parameters identification method, comprising:

S101: SCARA joint of robot kinetic model is established；

S102: parameter identification is carried out to improved friction model；

S103: substituting into kinetic model for the friction model after identification, rubs in the kinetic model except described Wipe the remainder linearisation outside model；

S104: based on the kinetic model after linearisation, observing matrix and restrictive condition is set, is changed to design Into the excitation track of Fourier formalism；

S105: being based on the excitation track, acquires corresponding data by experiment, to obtain dynamics ginseng to be identified Number；

S106: pass through kinetic parameter to be identified described in least squares identification.