Академический Документы
Профессиональный Документы
Культура Документы
Center for Robotics and Advanced Automation, Department of Electrical and Computer Engineering,
Oakland University, Rochester, USA
2
Electrical & Electronic Systems Engineering, Chrysler Group LLC, Auburn Hills, USA
Email: tporntha@oakland.edu, loh@oakland.edu
Received May 19, 2013; revised June 19, 2013; accepted June 26, 2013
Copyright 2013 Robert N. K. Loh et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
ABSTRACT
Electronic throttle control (ETC) system has worked its way to become a standard subsystem in most of the current
automobiles as it has contributed much to the improvement of fuel economy, emissions, drivability and safety. Precision
control of the subsystem, which consists of a dc motor driving a throttle plate, a pre-loaded return spring and a set of
gear train to regulate airflow into the engine, seems rather straightforward and yet complex. The difficulties lie in the
unknown system parameters, hard nonlinearity of the pre-loaded spring that pulls the throttle plate to its default position,
and friction, among others. In this paper, we extend our previous results obtained for the modeling, unknown system
parameters identification and control of a commercially available Boschs DV-E5 ETC system. Details of modeling and
parameters identification based on laboratory experiments, data analysis, and knowledge of the system are provided.
The parameters identification results were verified and validated by a real-time PID control implemented with an xPC
Target. A nonlinear control design was then proposed utilizing the input-output feedback linearization approach and
technique. In view of a recent massive auto recalls due to the controversial uncontrollable engine accelerations, the results of this paper may inspire further research interest on the drive-by-wire technology.
Keywords: ETC System; System Identification; Nonlinear Control; Input-Output Feedback Linearization; xPC
Target-Based Control System
1. Introduction
In the automotive industry, one of the fruitful technologies that have emerged from the increasing regulations in
terms of fuel economy, emission control, drivability and
safety is the drive-by-wire technology that creates the
electronic throttle control (ETC) system [1,2]. The system comprises of a throttle plate equipped with a preloaded spring and is driven by an electronic-controlled dc
motor to regulate airflow in the intake manifold. In modern vehicles, the engine control unit computes and maps
the throttle plates angle to many entries such as accelerator pedal position, engine speed, cruise control command, and so forth in order to achieve optimal air-fuel
mixtures in the combustion chambers, thereby maximizing fuel economy and minimizing emissions. The ETC
system, which responses to the prescribed reference from
the engine control unit must operate with fast transient
responses and precise control to regulate the throttle
plates angular position [3,4].
Copyright 2013 SciRes.
588
R. N. K. LOH
ET AL.
Throttle body
Throttle plate
Throttle plate
angle, (t)
Armature input
voltage, va(t)
DC motor
R. N. K. LOH
ia(t)
La
Ra
3. System Identification
nL > nm
nm
m(t)
+
_ M
eb(t)
va(t)
Return
Jm
L(t)
Tm(t)
KS
JL
Bm
nL
TL(t)
BL
dia t
Raia t K bm t va t .
dt
Substituting N m L
yields
La
and K b K b N
(1)
into (1)
dia t
Raia t K bL t va t .
dt
(2)
TL t
K mia t .
N
(3)
TL t J L L t BLL t
K s L t TPL T f sgn L
(4)
sgn L 0
if L 0,
if L 0,
(5)
if L 0.
(6)
589
ET AL.
(7)
La
ia t
Va
1 e t ,
Rt
(9)
R. N. K. LOH
590
3.5
Current (A)
2.5
1.5
Measured
Approximated
0.5
0
0
0.5
1.5
2.5
3.5
Voltage (V)
Voltage
1.6
ET AL.
10
1.2
1
100%
Current
Angular position
63%
0.6
0.4
0.2
0
-0.2
-0.001
0.001
0.002
0.003
0.004
0.8
1.4
Back emf
1
0.005
Time (s)
(10)
0.05
0.1
0.15
0.2
0.25
-10
0.35
0.3
Time (s)
read from Figure 4 when the current reached approximately 63% of its steady-state value was 1 ms; this yields
t . Therefore, the motor inductance was obtained as
La Rt 1.5 mH .
Angular position
2
0
Back emf
-2
-4
-6
Angular velocity
-8
-10
-12
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Time (s)
In order to measure the parameters in (11), the dc motor was energized by a slowly increasing armature current. This caused a slow rotation of the throttle plate. The
armature current and throttle plate angular position were
captured by the oscilloscope as shown in Figure 7.
It can clearly be seen from Figure 7 that a larger current was needed (approximately 1.7 A) to move the
throttle plate towards opening than towards closing (approximately 0.6 A). This was because the torque developed by the motor had to overcome the torques produced
by the return spring, K s L t , TPL , and the friction
torque, T f . Using the armature current data together
with the motor constant K m 0.4 obtained from the
previous experiment in (11), the load torque equation
becomes
TL t K s L t TPL T f ,
(12)
Angle
591
ET AL.
1.6
(0.240,1.571)
(0.524,1.571)
(0.808,1.571)
1.4
R. N. K. LOH
1.2
1
0.8
0.6
0.4
0.2
0
(0.112,0.096)
0
0.1
0.2
0.3
(0.396,0.096)
0.4
0.5
0.6
(0.680,0.096)
0.7
0.8
0.9
Current (A)
1.7 A
Current
0.5
0.6 A
10
12
Time (s)
14
0
16
1.5
R. N. K. LOH
592
(13)
where T fm represents static frictional torque of the motor. At any constant speed, m t 0 , therefore the frictional torque caused by viscous coefficient Bm is given
by
Tm t Bmm t T fm .
(14)
Optical sensor
ET AL.
4e
eb
b .
2 f 8 f
(15)
The procedures above were repeated with different arbitrary constant velocities and the corresponding eb and
K b were calculated. Figure 11 shows the result of the
FFT, and the calculated data of each test is shown in Table 1.
The overall average K b calculated from Table 1 was
0.0185 Vs/rad which is very close to the result of K b =
0.0193 Vs/rad obtained in Section 3.3 when the dc motor was not removed from the ETC unit. The method of
sensorless velocity measurement developed here was
judged to be more reliable than the method of Section 3.3,
and can be applied to the next parameters identification
described in the sequel.
3.5.2. Identification of Viscous Friction Coefficient
A series of step input voltages: 2, 4, 6, 8, 10 and 12 V,
6
16 pulses/revolution of optical sensor
4
2
0
-2
0.1
0.11
0.12
0.13
0.14
0.15
0.16
0.17
0.18
0.11
0.12
0.13
0.14
0.15
0.16
0.17
0.18
0.16
0.17
0.18
Voltage (V)
2.965
2.96
2.955
Current (A)
2.95
0.1
0.7
0.6
0.5
0.4
0.3
0.1
0.11
0.12
0.13
0.14
0.15
Time (s)
ENG
0.3
0.25
0.2
0.15
0.1
0.05
0
0
593
ET AL.
Amplitude
R. N. K. LOH
0
0.02
0.04
0.06
0.08
0.1
100
Time (s)
200
300
400
Frequency (Hz)
(a)
(b)
Figure 10. (a) Back emf voltage, and (b) its associated frequency.
Test #2
0.25
0.2
0.2
0.15
0.15
0.1
0.1
0.05
0.05
100
200
300
400
500
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
100
200
300
200
300
400
400
500
-2
0.5
1.5
1.5
1.5
1
0.5
0
-0.5
0.5
Time (s)
0
Frequency (Hz)
100
200
300
400
500
Frequency (Hz)
Average
back emf
eb (V)
Frequency f
(Hz)
0.8753
61.0426
0.01827
1.9009
129.41
0.018703
3.7113
255.77
0.018475
4.6501
321.69
0.018405
Average
2.7844
191.9782
0.0185
K b 4eb
(Vs/rad)
500
Frequency (Hz)
Test #4
0.6
100
Time (s)
Armature current (A)
Frequency (Hz)
Test #3
Test #1
0.25
R. N. K. LOH
594
Motor
torque Tm
(Nm)
95.2
74.80
0.405
0.0075
211.2
165.90
0.576
0.0106
334.6
262.76
0.688
0.0127
457.9
359.61
0.813
0.0150
10
580.0
455.51
0.903
0.0167
4. Parameters Verification
12
704.5
553.33
0.934
0.0172
The main objective of this section is to use the mathematical model described by (2) and (6) to verify the parameters identified in Table 3. Two methods were employed: SIMULINK simulation, and actual real-time
control implementation of the ETC system using an xPC
TargetBox. The xPC TargetBox is a real-time rapid prototyping device, which is widely used in both industries
emf (V)
-2
-4
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
0
0.02
-5
-10
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
100
0.015
0
-100
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
Torque (Nm)
Velocity (rad/s)
-6
0
Acceleration (rad/s 2)
Torque (Nm)
ET AL.
Measured
Approximated
0.01
-0.2
-0.4
0.02
0.04
0.06
0.08
0.1
0.12
Time (s)
0.14
0.16
0.18
T fm
0.2
0.005
100
200
0.0088 N m s rad.
300
400
500
600
Figure 14. Filtered back emf, velocity, acceleration and torque versus time.
Table 3. Results of ETC parameter identification.
Parameters and constants
Symbols
Identified values
Armature resistance
Ra
1.15
Armature measurement
circuit resistance
Rt
1.5
Armature inductance
La mH
1.5
K b V s rad
0.383
Km N m A
0.383
Spring constant
K s N m rad
0.087
TPL N m
0.396
Friction torque
Tf N m
0.284
Beq N m s rad
0.0088
J eq kg m
0.0021
ENG
R. N. K. LOH
y h x x1 .
y L2f h x Lg L f h x u
In this section, we applied the input-output feedback linearization technique [27-29] to design a nonlinear controller for the ETC system. For ease of analysis and reference, we define the following constants in the ETC
system described by (2) and (6):
(17)
Kb
Ra
TPL
1
k5
, k6
, k7
, k8 ,
J eq
La
La
La
k1 x1 k2 x2 k3 x3 k4 S x2 k5 ,
y L3f h x Lg L2f h x u
The resulting model can be expressed in a control-affine single-input single-output (SISO) nonlinear
state-space form as
x2
x1
0
x k x k x k x k S x k 0 u , (18)
2 1 1 2 2 3 3 4 2 5
k8
k6 x2 k7 x3
x3
Beq
Tf
K
K
k1 s , k2
, k3 m , k4
J eq
J eq
J eq
J eq
k2 k5 k2 k4 S x2 k4
dS x2
k3k8u
dt
where
Lif h x
Lif1h x
x
f x , Lg Lif1h x
Lif1h x
x
g x
x .
(22)
2 2
dt
1 e x2
g x
f x
(19)
595
ET AL.
Springs
0.1
Ks*(t)+Tpl
du /dt
Kd
Hard Limits
Km_bar
9
Kp
Step input
1
s
1/La
1
s
Ia
0.4
1
s
1/Jeq
12
Battery
(t)
1
s
(t)
Scope
Rt
Ki
Beq
(t)
emf
0.4
Kb_bar
Tf
Sign
ENG
R. N. K. LOH
596
ET AL.
1
T x
0
x
k1 k2 k4
Throttle Unit
Pedal
xPC TargetBox
Figure 16. Pictorial view of computer controlled ETC system with xPC Target.
(24)
z Az Bv,
y Cz ,
(25)
v
y b x D xu ,
0.8
(26)
0.6
0.4
0.2
Step input
Simulation
xPC Target
0.2
0.4
0.6
0.8
1.2
k2 k5 k2 k4 S x2 k4
1.4
1.6
1.8
e y y r z1 y r ,
yr ,
e
y
yr z2
e
y
yr z3
yr v yr .
Signum
= 0.5
= 1.0
= 3.0
-0.5
-1
-20
-15
-10
-5
0
5
Angualr velocity (rad/s)
10
15
20
z T x
x1
. (23)
x2
k1 x1 k2 x2 k3 x3 k4 S x2 k5
(28)
(29)
three and is equal to the dimension of the system. Therefore, the ETC system in (18) and (19) is fully linearizable
with the nonlinear transformation given by
(27)
0.5
0
dS x2
,
dt
D x k3k8 .
Time (s)
Magnitude
2 e x2
2
1 e x2
0
0 ,
k3
1.2
-0.2
0
1
(30)
Yr yr
y r
K R13 is determined such that Acl A BK is Hurwitz, that is, all eigenvalues of Acl lie in the open lefthalf complex plane. Finally, the linearizing feedback
control law can be obtained from (26) and is given by
u D 1 x b x v ,
(31)
where v(t) is given by (30), and b(x) and D(x) are given
in (27) and (28), respectively. Substituting (31) into (18)
ENG
R. N. K. LOH
y h x x1.
(32)
7. Conclusion
We presented the modeling and unknown system parameters identification of an industrial Bosch automotive
ETC system. The identification was based on a series of
laboratory experiments, measurements and data processing techniques. The validation of the identified parameters by comparing the simulation results with the
real-time implementation using an xPC Target under a
PID control scheme produced excellent results for this
particular ETC system which demonstrated the effectiveness and reliability of the parameters identification
techniques. We then proposed a nonlinear closed-loop
control system design based on the input-output linearization approach. The simulation results showed that the
nonlinear controller developed was capable of accom10
Angular velocity x (t), (rad/sec)
100
80
60
40
20
x 1(t)
Ref.
-20
0.6
1.2
Time (s)
1.8
0
-5
-10
0.6
1.2
Time (s)
1.8
2.4
0.6
1.2
Time (s)
1.8
2.4
10
-15
2.4
2
0
-2
-4
597
ET AL.
0.6
1.2
Time (s)
1.8
2.4
-5
-10
Figure 19. Simulation results of step reference input with different durations.
Copyright 2013 SciRes.
ENG
R. N. K. LOH
598
ET AL.
6
Angular velocity x (t), (rad/sec)
100
x 1(t)
80
Ref.
60
40
20
0
0.6
1.2
Time (s)
1.8
2
0
-2
-4
-6
-8
2.4
-20
2
1
0
-1
-2
0.6
1.2
Time (s)
1.8
0.6
1.2
Time (s)
1.8
2.4
0.6
1.2
Time (s)
1.8
2.4
4
2
0
-2
-4
-6
2.4
Figure 20. Simulation results of step plus ramp reference input with different durations.
3
Angular velocity x2(t), (rad/sec)
100
x 1(t)
Ref.
50
-50
-100
0.6
1.2
Time (s)
1.8
0
-1
-2
0.6
1.2
Time (s)
1.8
2.4
0.6
1.2
Time (s)
1.8
2.4
5
Armature Voltage u(t), (V)
-3
2.4
2.5
2
1.5
1
0.5
0
0.6
1.2
Time (s)
1.8
2.4
4
3
2
1
0
-1
R. N. K. LOH
by-wire technology.
REFERENCES
[1]
[2]
W. Huber, B. Lieberoth-Leden, W. Maisch and A. Reppich, Electronic Throttle Control, Automotive Engineering, Vol. 99, No. 6, 1991, pp. 15-18.
[3]
[4]
[5]
[6]
[7]
A. Contreras, I. Quiroz and C. C. de Wit, Further Results on Modeling and Identification of an Electronic
Throttle Body, Proceedings of 10th Mediterranean
Conference on Control and Automation, Lisbon, 2002.
[8]
[9]
ET AL.
599
R. N. K. LOH
600
Design, Wiley Interscience, Hoboken, 2003.
ET AL.
Springer-Verlag, London, 2001.
[31] C. Edwards and S. Spurgeon, Sliding Mode Control
Theory and Applicationsm Taylor & Francis, London,
1998.
ENG