Parallel robots are robots with closed kinematics structure in which the
links are connected by joints. Although the parallel robot has a complex
dynamic structure, and is difficult to design and control, but it has some
outstanding advantages over the serial robot: high load bearing capacity,
high rigidity due to configuration. They can perform complex operations and
operate with high accuracy. Therefore, study on the problem of dynamics
and control of the parallel robot in order to take advantage of it is a scientific
and practical matter.
23 trang |
Chia sẻ: thientruc20 | Lượt xem: 764 | Lượt tải: 0
Bạn đang xem trước 20 trang tài liệu Inverse dynamics and motion control of delta parallel robot, để xem tài liệu hoàn chỉnh bạn click vào nút DOWNLOAD ở trên
MINISTRY OF EDUCATION
AND TRAINING
VIETNAM ACADEMY OF SCIENCE
AND TECHNOLOGY
GRADUATE UNIVERSITY OF SCIENCE AND TECHNOLOGY
...***
NGUYEN DINH DUNG
INVERSE DYNAMICS AND MOTION CONTROL
OF DELTA PARALLEL ROBOT
Major: Engineering Mechanics
Code: 9 52 01 01
SUMMARY OF THE DOCTORAL THESIS
Hanoi – 2018
The thesis has been completed at Graduate University of Science and
Technology, Vietnam Academy of Science and Technology
Supervisor 1: Prof. Dr. Sc. Nguyen Van Khang
Supervisor 2: Assoc. Prof. Dr. Nguyen Quang Hoang
Reviewer 1: Prof. Dr. Dinh Van Phong
Reviewer 2: : Prof. Dr. Tran Van Tuan
Reviewer 3: : Assoc. Prof. Dr. Le Luong Tai
The thesis is defended to the thesis committee for the Doctoral Degree, at
Graduate University of Science and Technology - Vietnam Academy of
Science and Technology, on Date Month
Year 2018
Hardcopy of the thesis can be found at:
- Library of Graduate University of Science and Technology
- National Library of Vietnam
1
INTRODUCTION
The rationale for the thesis
Parallel robots are robots with closed kinematics structure in which the
links are connected by joints. Although the parallel robot has a complex
dynamic structure, and is difficult to design and control, but it has some
outstanding advantages over the serial robot: high load bearing capacity,
high rigidity due to configuration. They can perform complex operations and
operate with high accuracy. Therefore, study on the problem of dynamics
and control of the parallel robot in order to take advantage of it is a scientific
and practical matter.
2. The objective of the thesis
The objective of this thesis is to apply Lagrange equations with
multipliers to study dynamics and control of Delta parallel robots.
Particularly, mechanical model, mathematical model, and control algorithms
for Delta parallel robots are developed as a scientific basis for the research
and development of parallel Delta robots.
3. The object and the main content of the thesis
Research objects: Dynamics and control of two Delta parallel robots are
3RUS robots and 3PUS robots.
The main content of the thesis includes: Study of mathematical and
mechanical modeling problems, study of dynamics and control algorithms
for Delta parallel robot. The thesis does not study the problem of design and
manufacture of Delta parallel robots.
4. The outline of the thesis
The outline of the thesis contains Introduction, Four main chapters,
Conclusions and findings of the thesis.
Chapter 1: Overview of the study of dynamics and control of Delta
parallel robot in and outside the country is first presented. Since then, the
2
direction of the thesis has been selected to address scientific significance and
practical application.
Chapter 2: Presents the construction of mechanical models and
application of Lagrangian equations with multipliers to formulate
mathematical models for two Delta Parallel Robots. Each robot offers two
mechanical models for study and comparison.
Chapter 3: Presents some improvements in numerical methods to solve
the inverse kinematics and inverse dynamics of parallel robots. Inverse
kinematic problem is solved by applying the improved Newton-Raphson
method. Inverse dynamics problem is solved by reducing Lagrange
multipliers to calculate moments or driving forces in active joints.
Chapter 4: Presents tracking control of parallel robot manipulators
based on the mathematical model of parallel robots, which is a system of
differential – algebraic equations. The trajectory of serial robots described
by differential equations is often well studied. While the Delta parallel robot
trajectory is based on the mathematical model, the differential – algebraic
equations system is rarely studied. Control law such as PD control, PID
control, sliding mode control, sliding mode control using neural network are
studied in this chapter.
CHAPTER 1. OVERVIEW OF DYNAMICS AND CONTROL
PARALLEL ROBOT
1.1. Parallel robot
Parallel robots usually consist of a manipulator connected to a fixed
frame, driven in multiple parallel branches also called legs. The number of
legs is equal to the number of degrees of freedom. Each leg is controlled by
the actuator on a fixed frame or on the leg. Therefore, parallel robots are
sometimes referred to as platformed robots.
1.2. Comparison between Serial and Parallel Manipulators
3
Parallel robot has high rigidity and load bearing capacity due to load
sharing of each actuator operating in parallel. The accuracy of the position of
the parallel robot is high because there are no cumulative joint errors as the
serial robot. While kinematic chains create kinematic constraints and
workspace limitations, typical designs have low inertia characteristics. The
fields of parallel robot application include: CNC machine, high precision
machine, automation machine in semiconductor and high speed and high
acceleration electronics assembly industry. A comparison between parallel
and serial robots is given in the following table:
Table 1.1: Comparison between Serial and Parallel Manipulators
STT Features Serial robot Parallel robot
1 Accuracy Lower Higher
2 Workspace Large Small
3 Stiffness Low High
4 Payload Low High
5 Inertial Large Small
6 Speed Low High
7 Design/control
complexity
Simple complex
8 Singularity problem Some Abundant
1.3. Research on dynamics and control of parallel robots outside of
the country
1.3.1. Inverse dynamics of parallel robots
On the mechanical side, parallel robots are closed-loop multibody
system. Dynamic computation is essential to designing and improving the
control quality of parallel robots. The literature on the theory and calculation
method of robot dynamics is quite substantial [47, 73, 85-88, 96, 103]. The
methods of establishing the dynamic equations of closed-loop multibody
4
system are well documented in [88, 103]. The kinematics and dynamics
problems are then more specifically mentioned in the literature on parallel
robots [67, 96].
In the above studies on Delta parallel robots, the methods used to
establish equation of motion are Lagrange equations with multipliers, virtual
work principle, Newton-Euler equation, subsystem ... When establishing the
equation, the bar between the actuating link and the manipulator is modeled
with a uniform bar or with a zero-mass bar and two masses at the ends of the
bar. Up to now, there have been no comparative work on these two types of
models.
1.3.2. Tracking control of parallel robots
The documentation on robot control is very rich. There are various
approaches to controlling robots given by Spong and Vidyasagar [90],
Sciavicco and Siciliano [87]. However, these works are less focused on the
specific problems of parallel robots.
Recently, the works on improving the control quality of Delta robot was
also published quite a lot. These works develop control law based on the
equations of motion, which are obtained by simplifying the dynamics model
of each parallelogram by a zero-mass bar with two mass points at both ends.
Model linearization methods are used to establish simple control laws.
Hemici et al. [80-82] designed PID, H controllers based on linear models
to robustly control Delta robot. This model was also used by A. Mohsen [68]
to establish PD and PID control laws in combination with fuzzy supervision
to perform motion tracking control of manipulator.
These works use different controllers for the purpose of forcing the
movement of the manipulator to follow a desired trajectory. These
controllers partly meet the desired requirements. However, there is a lack of
5
comparative studies of controllers and recommendations on how to choose
an appropriate ones.
1.4. Studies in Vietnam
The research in Vietnam mainly focuses on solving the kinematics
problem, establishing the equation of motion and presenting the method of
solving the equations of motion. Control problems are little researched.
1.5 The research problem of the thesis
From the review and evaluation of the work that scientists have been
working on in Delta parallel robot, this thesis will investigate the following
issues:
Development of the solution for the inverse dynamics problem with the
aim of improving numerical accuracy.
Study and comparison of different dynamic models for a parallel robot,
the complexity of the models and their effect on the computational torque
moment. On that basis, it is advisable for the user to use a suitable model.
Design of direct control law based on differential – algebraic equations.
Research comparing the quality of the controllers using different
mechanical models.
Conclusions of chapter 1
Based on the results obtained from domestic and foreign researches, the
thesis has identified the need for in-depth research in order to improve the
quality of control for parallel robots, mechanical and mathematical modeling
and numerical algorithms for solving dynamic and control problems for two
parallel robots, 3RUS and 3PUS.
CHAPTER 2. BUILDING THE MECHANICAL MODEL AND
MATHEMATICAL MODEL FOR DELTA PARALLEL ROBOT
In this thesis, the new matrix form of the Lagrange equations with
multipliers [51] is used to establish the equation of motion of two parallel
6
robots, the 3RUS robot and the 3PUS robot. With the MAPLE or MATLAB
software, we obtain the analytic form of differential – algebraic equations
describing the movement of parallel robots.
2.1. Dynamic model of Delta parallel robot
2.1.1. Dynamic model of Delta parallel robot 3RUS
From realistic models of robots from Figure 2.1, it can be seen that the
parallelogram will make the kinematic and dynamic computation on the
robot quite complex. For simplicity we build two models of robot dynamics
based on real model as follows:
Figure 2.1: Delta parallel robot 3RUS
Model 1: The parallelogram mechanisms is modeled by a bar with a
uniformly distributed mass over the length of the bar. The mass and length
of the bars correspond to the mass and length of the parallelogram.
Model 2: The parallelogram mechanisms is modeled by a weightless
bar with a concentric mass at both ends, the mass of each bar end equals half
the mass of the parallelogram.
R1
m1,L1, I1
mp
m2, L2, I2
r
7
2.1.2. Dynamic model of 3PUS Delta parallel robot
Spatial 3PUS Delta robot is a variant of the 3RUS robot when
replacing rotary actuation joints linear actuation joints as shown in Figure
2.4. The 3PUS robot is also equipped with two dynamic models similar to
the 3RUS.
Figure 2.4: Delta parallel robot 3PUS
2.2. Establish equations of motion of the Delta parallel robot
Applying the new matrix form of the Lagrange equation with
multipliers [4, 51], the equation of motion of two 3RUS and 3PUS robots is
the differential - algebraic equations of the following general form:
, Ts M s s C s s s g s Φ s λ τ (2.20)
f s 0 (2.58)
2.3 Compare the equations of motion of robot models
From the equation of motion of model 1 and model 2 of each robot we
have the comparison table as follows:
8
Table 2.1. Compare the equations of motion of Models 1 and 2
Number of.. Model 1. Model 2.
Degree of freedom 3 3
generalized
coordinates
3x3 + 3 = 12 3 + 3 = 6
Constrained equations 9 3
Lagrangian multipliers 9 3
equations 21 9
Matrices M and C ( ), ( , )M M s C s s 0 ( ) , ( , ) 0constM s C s s
From Table 2.1 we find that the equation of motion of model 2 is
simpler and easier to establish than model 1, but the inertia effect is not
clear.
2.3. Conclusions of chapter 2
The establishment of analytical equation of the equation of motion of Delta
parallel robot is a very complex problem. Using the symbolic programming
technique, this thesis has achieved some new results as follows:
Using the new matrix form of equations Lagrange with multipliers [51],
the differential - algebraic equations describing the motion of the two kinds
Delta parallel robot (robot 3RUS and robot 3PUS) has established
analytically.
In addition to establishing equations of motion in view of rigid body,
the thesis also provides a simple equation for motion equation by replacing
the parallelogram mechanisms by two mass points. These mechanical
models are the basis for computational dynamics and control of parallel
robots 3RUS and 3PUS.
9
CHAPTER 3. NUMERICAL SIMULATION OF INVERSE
KINEMATICS AND INVERSE DYNAMICS FOR DELTA
PARALLEL ROBOT
Based on the explicit analytical form of the differential - algebraic equations
description of the motion of the Delta parallel robot set up in Chapter 2, this
chapter applies and develops numerical algorithms to solve the inverse
kinematic and inverse dynamic problem for parallel robots 3RUS and 3PUS.
3.1 Calculation of inverse kinematic parallel robot by improved
Newton-Raphson method
The constrained equations of robot are rewritten in vector form as follows:
( ) ( , )f s f q x 0 (3.1)
where: , , r n mf q x
Contents of the inverse dynamics problem: Given the motion law of the
manipulator, it is necessary to find the law of motion of the driving joints.
Here, we will present an improved Newton-Raphson method [4] to solve the
inverse kinematic problem:
Step 1: Correct the increment of the vector of generalized coordinates at
time t0 = 0. First, we can determine the approximate vector 0q by drawing
method (or experiment). Then apply Newton - Raphson methods to find a
better solution of 0q from nonlinear equations (3.1).
Step 2: Correct the increment of of the vector of generalized coordinates
at time tk+1. The approximate initial value of qk+1 is approximated by the
formula:
2
1
1
( )
2k k k k
t tq q q q
In the robot kinematics computation [87], the infinitesimals of order
n≥2 are often neglected in the initial approximation of Newton-Raphson. In
this thesis, we take into account the second order infinitesimals, neglecting
10
the infinitesimals of order 3 and taking the formula (3.14) as the
approximation of the original Newton-Raphson loop.
After each step of calculating the coordinates of the joints using the
improved Newton-Raphson method, the generalized velocity and
acceleration of the joints are calculated by the following formulas:
1
q x
q J J x (3.4)
1q q x xq J J q J x J x (3.6)
3.2 Numerical method for solving the inverse dynamics problem of
parallel robots
3.2.1 Inverse dynamics problem
The general equations of motion of the robot is as follows:
T
sM(s)s +C(s,s)s +g(s)+Φ (s) (3.20)
( )f s 0 (3.21)
Let f
a q be the vector of coordinates of active joints,
rz is the
vector of redundant coordinates (including passive coordinates and end-
effector coordinates). Symbol:
, , , , ,
TT T n f r
a a n f r s q z s q z
The inverse dynamics problem of the parallel robot is expressed as: The
equation of motion of the robot is known as in (3.20), (3.21), given the
motion law of the operation , mt x x x . Determine fa τ the
driving momen / force required to produce the desired motion.
3.2.2 Solving the inverse dynamics problem by eliminating the Lagrange
multipliers [4]
Through the inverse kinematic with the given trajectory of the mobile
platform center, we have found the vector , , t t ts s s . From this, mass
matrices, centrifugal inertia and Coriolis matrices, matrices sΦ , as well as
the vector g(s) have been completely determined. Thus, Equation (3.20) is a
linear algebraic equation with unknown driving torque vector aτ
and
11
Lagrange multipliers λ with equal numbers of equations and numbers.
Thus, we can directly solve this system of equations and then separate the
resulting momen.
In this thesis, we will not directly solve equation (3.20) but try to
eliminate Lagrange multipliers λ , transforming the system of differential -
algebraic equations (3.20), (3.21) into the system of equations of only
unknowns of only joint moments aτ as follows:
We put the symbol [4, 47]:
1( , )a
z q
s
E
R R q z
Φ Φ
(3.24)
where E is the unit matrix size f f and ,
a
z a
f f
Φ Φ
z q
Left multiplying both sides of (3.20) by matrix T sR and simplying it
yields
,T T T as s sR M s s R C s s s R g s τ (3.29)
The expression in the left-hand side of equation (3.29) are known from the
results of the inverse kinematics problem. Thus, active joint moments are
calculated according to this equation.
3.3. Numerical simulation of inverse kinematics and inverse dynamics of
Delta parllel robot
3.3.1. Numerical simulation of 3RUS inverse kinematics of robot
To evaluate the correctness of algorithms and calculations of the thesis,
we computed the inverse dynamics problem of 3RUS robot with the
DELTA-IMECH program developed based on MATLAB software. For
comparison, the robot parameter data and manipulation motion are given in
[61] of Y. Li and Q. Xu.
Using the DELTA-IMECH program we obtain the results of the
numerical simulation of inverse kinematics and have the following
comparison table:
12
The results of the thesis The results of work [61]
Figure 3.11: Comparison of the results of the inverse kinematic problems against
the literature [61]
0 0.5 1 1.5 2
20
40
60
80
100
t[s]
[d
e
g
re
e
]
Joint
1
Joint
2
Joint
3
0 0.5 1 1.5 2
-2
-1
0
1
2
t[s]
[r
a
d
/s
]
0 0.5 1 1.5 2
-4
-2
0
2
4
6
8
t[s]
[r
a
d
/s
2
]
13
Comment: Figure 3.11 shows that the results of the inverse kinematics of
the thesis are consistent with the results of the paper [61].
3.3.1. Numerical simulation of inverse dynamics of robot 3PUS
The robot parameter data and movement of the manipulator as follows:
1 20.242, 0.16, 0.029( ), 0.12, 2 0.15, 0.2(kg)PL R r m m m m
Numerical simulation results were computed based on models 1 and 2 of the
3PUS robot using the DELTA-IMECH program.
Model 1 Model 1
T = 1 (s), (Fast motion manipulator)
T = 10 (s), (Slow motion manipulator)
Figure 3.22: Results of numerical simulation inverse dynamics robot 3PUS
2 2
0.05cos ; 0.05sin ; 0.5 ( )P P Px t y t z m
T T
0 0.2 0.4 0.6 0.8 1
-6
-5.5
-5
-4.5
-4
t[s]
[N
]
0 0.2 0.4 0.6 0.8 1
-5.2
-5
-4.8
-4.6
-4.4
t[s]
[N
]
0 2 4 6 8 10
-7
-6
-5
-4
-3
t[s]
[N
]
0 2 4 6 8 10
-7
-6
-5
-4
-3
t[s]
[N
]
14
Comment: When motion of the manipulator is fast,