A High Accuracy Temperature Control System Based
on ARM9
Xiaofang Li
Yuntao Yu
College of Automation and Electronic Engineering
Qingdao University of Science and Technology
Qingdao, Shandong Province, China
lxfzlh@163.com
College of Automation and Electronic Engineering
Qingdao University of Science and Technology
Qingdao, Shandong Province, China
skyyuyuntao@163.com
Abstract—In the industry and agriculture, a temperature
object often has many complex characteristics such as
nonlinear, large time delay, time-variable, strong
coupling and so on. We introduce a kind of the fuzzy
self-tuning PID temperature control system based on
ARM9.Experiment results are shown that this system
has high accuracy, rapid respond ,good real-time
capability and good stability performance.
Keywords- temperature control; fuzzy self-tuning PID; ARM9;
embeded linux
I.
INTRODUCTION
The system mainly by the temperature measurement
component, the ARM9 controller and the display transmission
unit three parts to be composed. System structure as shown in
Figure 1.
LCD monitor
Clock and reset
circuit
FLASH
s3c2440
Power output
circuit
SDRAM
II.
DESIGN OF SYSTEMATICAL HARDWARE
A. Microprocessor S3C2440A
The S3C2440A offers outstanding features with its CPU
core, a 16/32-bit ARM920T RISC processor designed by
Advanced RISC Machines, Ltd .The ARM920T implements
MMU, AMBA BUS, and Harvard cache architecture with
separate 16KB instruction and 16KB data caches, each with an
8-word line length [1]. A system of NAND FLASH start is
adopted. NAND FLASH memory expansion options
manufactured by Samsung K9F1208, single capacity of 64M ×
8bit (64MB), operating voltage 2.7 ~ 3.6V, 8-bit data width,
data protection with hardware support automatic power boot
functionality. SDRAM system selects HY57V561620T,
monolithic storage capacity of
4x4Mx16 bit (32MB),
operating voltage (3.3 ± 0.3) V, 16-bit data width. This system
selects HY57V561620T to construct 32 SDRAM memory
system, altogether the 64MB SDRAM space, may satisfy the
embedded operating system and each kind of relatively
complex function movement request.
B. Temperature gathering unit
This system uses the department which American
Semiconductor DALLAS Corporation recent years promoted
to draw up a series of single-bus temperature sensor chip
DS18B20.
The chip has the following characteristics：
1）Unique 1-Wire interface requires only one port pin for
communication
Keyboard
input
Figure 1. Overall system block diagram
2）Can be powered from data line. Power supply range is
3.0V to 5.5V
3 ） Measures temperatures from -55°C to +125°C.
Fahrenheit equivalent is -67°F to +257°F
4）±0.5°C accuracy from -10°C to +85°C
The principle of the system is as following:
Firstly we set the temperature value through the keyboard.
Then the ARM9 controller which utilizes the temperature senso
r to gather the temperature signal and output the PWM wave to
control the power module through the fuzzy PID control modul
e achieve the heating and cooling temperature. The last we disp
lay the temperature through the LCD.
Information is sent to/from the DS18B20 over a 1-Wire
interface, so that only one wire (and ground) needs to be
connected from a central microprocessor to a DS18B20.
Power for reading, writing, and performing temperature
conversions can be derived from the data line itself with no
need for an external power source [2]. The DS18B20 Digital
Thermometer provides 9 to 12-bit (configurable) temperature
readings which indicate the temperature of the device. This
system uses the GPB7 pin to actuate DS18B20.
C. Keyboard and LCD display unit
System uses the SPI interface control chip ZLG7289 the
SPI interface to connect with S3C2440A. ZLG7289’s row lines
R [2:0] and column lines C [7:0] constitute the matrix of the
keyboard, while inside the chip can be completed automatically
scanning, decoding, to dithering and other tasks.
The S3C2440A already integrated the LCD controller,
therefore may very only then control each type the LCD
screen, for example: STN and TFT screen. System uses a
Samsung LCD screen LTS350Q1, which is composed a TFTLCD module, a driver circuit and a back-light unit. The
resolution of a 3.5" contains 320RGB*240 dots and can display
up to 16.7M colors [3]. When that can do a connection fact of a
drive circuit of S3C2440A touch panel directly, it's possible to
touch the location and get it by sampling of an ADC electric
circuit CPU has built-in directly. Keyboard and LCD
connection diagram as shown in Figure 2 and Figure3.
SPIMOSI1
MOSI
SPIMISO1
MISO
SPICLK1
SCK
nSS1
nSS
meet the requirements of system reliability. Additionally Linux
was easy to transplant, and a cut core offered a good
technological prop for the systematical development by which
small efficient source cord is openhearted and has a lot of
developers [4].
System development intersects first, and information
bootloader establishes to compile the environment, and
transplants OS, loads with a file system and develops figure
interface, and application is compiled at the end [5]. System is
based on QT / E of the graphical user interface. Q/E has
continued Qt in tabletop system's all functions, rich API
interface and component-based programming model makes
embedded Linux application development system is more
convenient. System program flow as shown in Figure4.
Start
System initialization
Read settings
EINT1
nKEY
R[2:0]
S3C2440A
VLINE
VCLK
VM
VD[7:0]
S3C2440A
Setting mode
Default mode
Gathers and demonstrates
various
test
points
temperature
FLM
CP
LOAD
Transfers the fuzzy
control subroutine
DF
D[7:0]
LTS350Q1
Figure 3. LCD connection diagram
III.
N
Y
C[7:0]
ZLG7289
Figure 2. Keyboard connection diagram
VFRAME
The system has established?
DESIGN OF SYSTEMATICAL SOFTWARE
A. Embedded linux operating system
Linux operating system has a complete TCP / IP protocol,
good stability and real-time, good intelligent control system to
Output control signals
Figure 4. system program flows
B. Fuzzy PID control algorithm
A fuzzy self-tuning PID controller's core is to find fuzzy
relation between PID controller’s three parameter KP, KI, KD
and the erroneous absolute value |E| and erroneous rate of
change absolute value |Ec|. Through continuous testing in the
operation of | E | and | Ec |, according to the fuzzy control rules
to the three parameters on-line modified to meet different | E |
and | Ec |, so that the controlled object has good dynamic and
static performance [6].
1) The establishment of input and output variables
Based on the system’s above analysis, takes erroneous E
and erroneous rate of change Ec as the fuzzy controller's input,
and PID controller's three parameter KP, KI, and KD as the
output[7].
2) Input and output variables of the fuzzy language to
describe
Set of input variables | E | and | Ec | fuzzy sets as {NB, NM,
NS, ZO, PS, PM, PB}. Erroneous E and erroneous rate of
change Ec arrives inside the area of (-3, 3). Similarly, design
output KP, KI, KD fuzzy subset is {ZO, PS, PM, PB}, and its
quantification to region in (0, 3) [8]. Input and output variables
of the membership function curves are shown in Figure5 and
Figure6.
1.0 NB
NM
NS
Z
PS
PM
PB
larger system response in order to avoid the overshoot, the
integral action should be limited, usually taken KI = 0.
When | E | in the middle size, for the system response has a
smaller overshoot, KP should be made smaller. In this case, the
KD value is big to system response's influence, and the KI
value must be suitable.
When | E | is small, as the system has good stability, KP
and KD should be made bigger.
Meanwhile to avoid the system in designs nearby the
definite value to present the vibration, KD value choice basis
|Ec| determined. When | Ec | value is lesser, KD choose big
some; When | Ec | value is bigger, take smaller value KD. KD
is usually moderate in size.
4) Fuzzy control rule table
According to the above the PID parameters tuning
principles and expert experience, can list the output variables
KP, KI, KD control rules in Table 1 to Table 3.
0.8
0.6
TABLE I.
KP CONTROL RULE TABLE
TABLE II.
KI CONTROL RULE TABLE
0.4
0.2
0
.
-3
-2
-1
0
1
2
3
Figure5. Input E, Ec membership function curve
10
ZD
PS
PM
PB
8
6
4
2
0
-3
-2
-1
0
1
2
3
Figure6.Output KP, KI, KD membership function curve
3) The principle of PID parameter tuning
Relevant professional literature summarizes the system is in
process control for different | E | | Ec |, parameters KP, KI, KD
setting of principles [9]:
When | E | is large, as the system has good tracking
performance, should take a larger Kp and smaller KD, while a
Fuzzy Reasoning
KP
SV
+
_
KI
KD
E
PID controller
PV
Object
dE/dt
EC
Figure7. Fuzzy self-tuning control PID parameter block diagram
TABLE III.
control flexible, fast response and robust performance is strong,
and has the advantages of high precision of the classic PID
control characteristics, in transition process time, maximum
overshoot, etc, are better than the classical PID control. It’s a
high precision and strong stability temperature control system
that can be applied in different control object under the
different environment. It provides an effective method solve
the nonlinear, time-varying, large time delay of control for
many industrial processes.
KD CONTROL RULE TABLE
REFERENCES
[1]
Taking the conventional PID control as the foundation, the
system uses fuzzy reasoning thought according to different E
and Ec to tune the PID parameter on line [10]. According to the
idea, this control system consists of two parts, namely the
routine PID control section and fuzzy reasoning parameters
calibration parts [11]. System structure diagram as shown in
Figure7.
The PID parameters will be brought into the following
formula:
k
u(k )  K P e(k )  K I T  e( j )  K D e(k ) / T
（1）
j 1
In the formula, u (k) is output in k time in the system, and e
(k) is the deviation of k time, and T is the cycle of sampling.
Outputs u (k) to be multiplied by corresponding proportionality
factor Ku then to obtain precise output U.
U  K u u(k )
（2）
IV.
CONCLUSION
This system based on the most popular single bus
temperature sensor, high-performance ARM9 processor and
the embedded Linux operating system design, design and
implements a high precision temperature controller. This fuzzy
self-tuning PID temperature control system not only with fuzzy
Samsung Electronics Co.Ltd.S3C2440A 32-BIT RISC Microprocessor
User S Manual ［ Z ］ .Korea: Revision 0.12, Samsung Electronics
Co.Ltd, 2004.
[2] Yong Hoon Choil, Woo Kwon1,Heung Nam Kim1,“Code generation for
Linux device driver,”Computer Society,2006(3)：14-17..
[3] Chia-Feng Juang and Jung-Shing chen, “A recurrent neural fuzzy
network controller for a temperature control system,”Fuzzy
Systems,2003.The 12th IEEE International Conference on Volume
1,25-28 May 2003.
[4] Lisboa,Portugal, “How to build a Timely Computing Base using RealTime Linux”, IEEE,2008:2810-2813.
[5] Yong Hoon Choil, Woo Kwon1,Heung Nam Kim1,“Code generation for
Linux device driver,”Computer Society,2006(3)：14-17..
[6] Aly,A.A and El-Lail, “Fuzzy Temperature Control of A Thermoelectric
Cooler,”Industrial Technology,2006.ICIT 2006.IEEE Intrenational
Conference on 15-17 Dec.2006.
[7] Woosung Choi,Woojong and Sangchul,“Development of automatic
temperature
control
system
in
blast
furnace,”
SICEICASE,2006.International Joint Conference.
[8] Huzmezan, M.,Gough,B. and Kovac,S,“Advanced control of batch
reactor temperature,” American Control Conference,2002.Proceedings
of the 2002 Volume 2,8-10 May 2002.
[9] I.J.Gyongy,D.W.Clarke,“On the automatic tuning ang adaptation of PID
controllers,” Control Engineering Practice.2006.14.149-163.
[10] Zhi-Wei Woo,“A PID type fuzzy controller with self-tuning scaling
factors,” Fuzzy Sets and Systems,2000.115,321-326.
[11] Hasaan B.Kazemian,“Comparative study of a leraning fuzzy PID
controller ang a self-tuning controller,” ISA Transactions.2001.40.245253.