International Journal on Electrical Engineering and Informatics - Volume 8, Number 1, March 2016
Automatic Gain Fuzzy Logic Controller for Pulse Radar Receiver System
Eko Joni Pristianto1 and Pranoto Hidaya Rusmin2
1
Research Center for Electronics and Telecommunication, Indonesian Institute of Sciences, Indonesia
2
School of Electrical Engineering and Informatics, Institut Teknologi Bandung, Indonesia
1
ekoj001@lipi.go.id, 2pranoto@stei.itb.ac.id
Abstract: Pulse radar is a device for emitting short pulse with high power and receiving echo
signal from the target. At receiver side, pulse radar has an Automatic Gain Control (AGC)
module to adjust receiver sensitivity and get the best amplitude level for the next process. The
AGC type used in this research is HMC992LP5E. In previous research, gain settings used a
classical closed loop AGC system. The disadvantage of this system is slow response of the
AGC to change the input signal. As the result, it is difficult to distinguish between noise and
target signal and differentiate amplitude level signal for different targets on radar system
application. This paper explains the design of Automatic Gain Control (AGC) system using
HMC992LP5E and fuzzy logic controller. The gain value of AGC is determined by input and
output signal amplitude level of AGC and also the maximum range of radar targets. AGC gain
value setting is provided by a certain voltage value at the pin VCTRL. VCTRL value is
calculated by a 32-bit microcontroller using fuzzy logic algorithms. This controller will
generate more dynamic value of AGC gain. The AGC output signal amplitude level can be
determined based on fuzzy logic algorithms. So that, the target signal captured by the pulse
radar receiver system will be more easily identified and can be used for other applications.
Keywords: pulse
microcontroller.
radar,
automatic
gain
control,
HMC992LP5E,
fuzzy
controller,
1. Introduction
Pulse radar is a device that emits high power short wave and in some periods will receive
echo signal. Its receiver part receive back the reflected electromagnetic wave from object
signal that was detected by radar through antenna reflector. Generally, receiver has ability to
filter the echo signal and make it suitable with the preferred detection, amplify weak object
signal and pass it to signal processing system, and display the final data to the display system
[1].
On the receiver part, there is an Automatic Gain Control Module (AGC). AGC module
works to adjust receiving sensitivity and gain best signal amplitude before further process [2].
In this research used HMC992LP5E AGC type, which is produced by Hittite Microwave
Coorporation. This AGC is a Intermediate Frequency (IF) analog signal Variable Gain
Amplifier (VGA) controller. It is consisting of two identical variable attenuators combined
with Monolithic Microwave Integrated Circuit (MMIC) amplifier that works on frequency
50MHz to 800MHz. The gain controlled value has ranges from -10 dB to 40 dB [3].
In the classical closed loop AGC system, the amplification setting has implemented by
setting the VGA, which is integrated with the logarithmic detector (log detector for short).
Then, the amplitude of AGC output signal adjust by giving a voltage set value VSET for set
point ranging from 0.2 to 1.2 VDC. In previous research, the AGC gain setting was
implemented conventionally by giving a constant voltage for VSET, so that AGC output signal
amplitude will be fixed to the given set point value [4]. Classical AGC has several drawbacks.
It has slow response compared to fast change on input signal. This system result similar
amplitude for different targets or noise which causes difficulty to determine different target or
distinguish noise from target signal.
Received: January 10th, 2016. Accepted: March 17th, 2016
DOI: 10.15676/ijeei.2016.8.1.5
62
Eko Joni Pristianto, et al.
Figure 1. RCS on stealth aircraft Boeing UCAV X-45 (S Band)
Every pulse radar system target has Radar Cross Section (RCS), a ratio of backscatter
density in the direction of the radar (from target) to the power density of signal. This value is a
density parameter of compactness of the detected object [2]. As an example, Figure 1 shows
RCS value of a Boeing UCAV X-45 [5]. The following Table 1 shows some RCS values of
different objects.
Table 1. RCS value of different objects
According to the Table 1, an aircraft carrier has greatest RCS value which is 100000 m2 (50
dBsm), followed by cruiser, large airline, medium airliner, large fighter, Man, until insects with
RCS value of -60dBsm [5]. The RCS value, which is target signal amplitude, will be received
by radar receiver system. Thus, to obtain variety of target signal, the output amplitude of AGC
cannot be fixed all the time.
With this potential advantage, an AGC HMC992LP5E controller using fuzzy logic needs to
be designed. In previous researches, AGC fuzzy controller has been developed in many
systems such as the application of fuzzy logic to improve the performance of AGC circuit. For
this case, AGC simulation using amplifier circuit [6]. The AGC based on fuzzy algorithm is
applied for Single Side Band (SSB) radio communication system [7]. AGC control system
based on fuzzy logic for Erbium Doped Fibre Amplifier (EDFAs) [8]. And, the fuzzy control
for AGC microphone in Autonomous Support System For Elderly [9]. The following Table 2
describes previous results from input-process-output perspectives.
63
Automatic Gain Fuzzy Logic Controller for Pulse Radar Receiver System
No
1
2
3
4
Table 2. Recent Researches on AGC Fuzzy
Research
Features
AGC simulation using Input: two inputs: input dan output signal as ac signal
amplifier circuit
with specific frequency.
Process: one module fuzzy but the fuzzy inference system
is not explained.
Output: the Vagc voltage to determine op-amp gain on
the transistor basis.
The AGCbased on Input: two inputs: audio signal strength and its derivative.
fuzzy algorithm is Process: one module fuzzy using mamdani inference
applied for Single system.
Side Band (SSB) radio Output: the signal gain factor
communication
system
AGC control system Input: two inputs, signal power and wavelength
based on fuzzy logic Process: one module fuzzy using mamdani inference
for Erbium Doped fuzzy model
Fibre
Amplifier Output: current of the pump laser
(EDFAs)
Fuzzy control for Input: one input, Level of Noise
AGC microphone in Proses: one module fuzzy but the fuzzy inference system
Autonomous Support is not explained.
System For Elderly
Output: Microphone sensitivity.
In this research, the value of AGC gain is determined according to input signal amplitude,
AGC output, and maximum range of radar target. The setting of AGC HMC992LP5E gain
value can be done by giving a certain voltage value to pin VCTRL. Calculation of VCTRL
value using fuzzy algorithm will be executed by 32 bit microcontroller. This control system
generate dynamic AGC gain value and AGC output signal based on the fuzzy logic algorithm.
2. Classic closed loop AGC HMC992LP5E system
HMC992LP5E can be configured as AGC amplifier by VGA core and log detector in single
chip. The following Figure 2 shows a block diagram and classic closed loop
AGCHMC992LP5E circuit.
In this configuration, input signal is amplified by VGA. Then, output of VGA core is fed
back to log detector input RFDETIN (pin 16) through an external cupler to decrease the
maximum or minimum amplitude of VGA output such that the value in the dynamic range of
log detector. Log detector generate voltage at DETOUT (pin 13) proportional to VGA output
power amplitude. Output pin DETOUT with high impedance is connected to gain controller on
pin VCTRL or VGA to form AGC loop. Pin VSET has output power constant which is the set
point value from AGC amplifier. VSET value is a negative feedback, so the VGA gain
automatically will be adjusted to make AGC output signal power remains constant. Output
signal power is determined by value of VSET AGC, regardless of input signal variance.
Figure 3 shows the correlation between input and output power at different VSET values in
classic AGC configuration [3].
64
Eko Joni Pristianto, et al.
INPUT
RFIN
Automatic Gain Control
AGC HMC992LP5E RFout
50 - 800 MHz
VSET
Coupler OUTPUT
-10 dB
RF det IN
Variable
voltage
RF signal
DC signal
(a)
(b)
Figure 2. (a) Block diagram, (b) loop AGC HMC992LP5E schematic with two attenuators and
two amplifiers configuration [3].
Figure 3. The Correlation between input and output power at different VSET values in closed
loop AGC configuration
Input power under and over the range value of attenuation will be saturated, so that the
output is a linier function of input.
65
Automatic Gain Fuzzy Logic Controller for Pulse Radar Receiver System
3. Proposed AGC System using Fuzzy Logic Controller
This fuzzy logic controller makes the gain value of AGC fluctuate. While, the value of gain
on classical closed loop AGC system is static because it is only has one gain value. In general,
this algorithm can only be applied in receiver system of non pulse compression Radar and it is
not yet tested for other type of Radar such as Frequency-Modulated Continuous-Wave
(FMCW) radar. The block diagram of AGC HMC993LP5E gain control using fuzzy logic can
be seen on below Figure 4.
Splitter 600 Mhz
(Input)
RFIN
Automatic Gain Control
AGC HMC992LP5E RFout
50 - 800 MHz
Splitter 600 Mhz
(output)
OUTPUT
signal
VCTRL
y
RF Power Detector
(Input)
u1
Microcontroller
STM32F401RE
u2
RF Power Detector
(Output)
Figure 4. AGC HMC993LP5E gain control using fuzzy logic
There are two two-way splitters, which split input signal into AGC input port and RF power
detector input port – which is used to read signal amplitude to the AGC. Output splitter split
AGC output signal into RF power detector port and system output port which is used to read
signal amplitude the AGC.
RF power detector convert RF signal into DC signal. Then, the DC signal will be the input
for microcontroller’s ADC and fuzzy logic variable. Where, the fuzzy logic controller inputs
are detection signals, u1and u2. The microcontroller process the value of u1 and u2 by fuzzy
algorithm and generate PWM output signal. This PWM signal is connected to VCTRL pin at
the AGC. The detailed circuit of system on Figure 4 is shown on Figure 5.
Figure 5. Schematic of AGC HMC992LP5E using fuzzy logic controller
To apply this method, log detector circuit has to be deactivated by giving 0V to DETEN pin
(no 18). This configuration uses two attenuators and two amplifiers.
4. AGC Signal and System
AGC HMC992LP5E fuzzy logic gain control is not only using input and output signal
amplitude. Other variables outside AGC system can also be used as fuzzy input value to
determine VCTRL value. In this research, AGC gain control method with fuzzy logic will be
66
Eko Joni Pristianto, et al.
applied to control pulse radar target signal amplitude. A fuzzy input variable which is a
maximum range of desired radar target will be added. Figure 6 shows a block diagram of AGC
application with fuzzy control on pulse radar target signal amplitude control.
1
Splitter
600 Mhz
(Input)
Automatic Gain Control
AGC HMC992LP5E
50 - 800 MHz
2
6
Splitter
600 Mhz
(output)
12
7
3
10
RF signal
DC signal
PWM signal
Data
PWM to
Analog
9
RF Power
Detector
(Input)
4
8
Microcontroller
STM32F401RE
5
PC
RF Power
Detector
(Output)
11
LCD
Figure 6. AGC HMC992LP5E using fuzzy logic controller on pulse radar target amplitude
control
On the first process, signal 1 in the form of continuous wave will be generated by signal
generator. Signal 1 pass to a input splitter to generate signal 2 and signal 3; each will be
attenuated by -3 dB. Then, signal 3 pass to RF power detector input. On RF power detector
input, signal 3 will be converted to DC signal to generate signal 4. Signal 4 is connected to
ADC 0 input pin of microcontroller. Then, ADC data 0 is converted again to input signal
which will be the value of fuzzy input, u1.
On the second process, microcontroller receive data signal in serial format from personal
computer. This data is the maximum detected range value. This value become the value of
fuzzy input, u2. Then, microcontroller calculate the first fuzzy algorithm with u1 and u2 as input
variables to generate first fuzzy output, signal 9, which is a PWM signal. First, this PWM
signal is converted to DC signal (signal 10) and will be the input for VCTRL AGC.
On the third process, after AGC gets the VCTRL value from microcontroller, the gain
adjustment will be executed by HMC992LP5E and signal 6 is generated. Signal 6 will be
processed by similar method to the first process to generate signal 12 and signal 7. Signal 12 is
AGC output signal with -3 dB attenuation and signal 7 will generate the second fuzzy input
value, u3 as feedback from system. Microcontroller calculate the second fuzzy algorithm with
y1 and u3 as input variables. The result update signal 9. All systems shown on above Figure 6,
operate on time domain and there is no frequency change. In this research, the frequency of
input signal is 600 MHz.
5. Radar Target Simulator
This module generate signal that represents target which will be received by pulse radar.
The generated target signal refers to the following radar equations (1).
𝑆𝑚𝑖𝑛 =
𝑃𝑡 𝐺𝐴𝑒 𝜎
(1)
(4𝜋)2 (𝑅𝑚𝑎𝑥 )4
Where
Rmax = Maximum range.
Pt = Transmitter amplitude.
G = Maximum gain antenna.
Ae = Effective area of the receiving antenna.
67
Automatic Gain Fuzzy Logic Controller for Pulse Radar Receiver System
σ = Radar cross section.
Smin = Minimum detectable signal.
Smin is amplitude of target with the change of Rmax value. The values of Pt, G, Ae, and σ are
constant. The block diagram of pulse radar target signal generator is shown in Figure 7 below.
Signal
Generator
Digital
Attenuator
INPUT
OUTPUT
AGC
RF SWITCH
z
Microcontroller
TX
Figure
7. Pulse Radar Target Signal Generator
RF SWITCH
R F S W ITC H
TX
IF = 60 0MHz
TX of digital attenuator by assigning 6 bit
The Microcontroller control the attenuation value
biner configuration attenuation,
which has value ranging from -0.5 dB to -31.5 dB, that is
IF = 6 0 0 M H z
TX
randomly generated within certain period. The module output is a 600 MHz signal that has
TO R
been modulated with different amplitude level. Figure 8 shows several waveform of pulse radar
PULSE MODULATOR
P U L S E M O D U L Atarget
T O R signal generator.
DULATOR
R F S W IT C H
R
0 =M 6H 0 z 0 M
H
F
S W
R
F
IT S
C W
H
R
IF
C
R
H
T
I
X
FT
X =
P
P
W
P
I
PM
I
T
C
F
F
6H
0I
FzT
M
L
S
O
0
X =M
P ED
F
P
R
F
U
F
I
M
R P
P
TT
R
P
I
P
U
XX
P
M
F
W
=M
6H
P RF
F
I 0 Fz
0
AP
ED T U
UM
L
L OS
M
P
TT
R P
P
F W
M
I
P
F W
XX
P
M
F
0
F
I
z
0
U OM L
L RSO A
T
S
C
M
E DT
WH R
H
U
O
M
IF T
A
U
FT
X =
L
X
I
S
6
I
T
RP
M
FW
RP
P
FW
W
P
P E
U
M
L
S
O
I0 FT
0 X
=M
6H
0I
P
W
P
IF
E U ML S
O ED U
M LO A DT UO LR A T O
U
UM
FzT
X =M
L
L OS
AP
ED T U
6H
I 0 Fz
0
z
U OM L
L RSO A
P RF
= M
F
6 H
S
0
E DT
z
W R
F
FI
T
SC
S W
R
T
R
O
P
WH
I
T
C
H
X
T
X
T
X
T
X
R
R
F
P
LR O
A
H
S
C
D
T
R
U O
F
L R
P
A
T
R
R
CS
W
HR
FI
T
F
P
F
R
P
R
F
P
F
R
P
R
F
P
F
R
P
R
F
F
P RF
P RF
T
IF T
R
F
O
z
WH R
R
P
W
M
F
I
T
W
U
O
M
0
A
F
Signal 2PW
TX PRF
SC
WH
I
T
C
X
T
X
T
X
T
X
H
Figure 8. Several waveform of pulse radar target signal generator
Signal 1 is an input signal coming from generator. Signal 2 is a modulated signal with
different amplitude. This signal represents the radar target [4].
P U LS E M O DU LA TO R
R
L R
R
P
P ED
0
U O
P
F
R
H
W
HR
T
D
W
P
CS
z
LR O
XX
P
P
F
H
0 =M 6H 0 z 0 M
W R
6 H
PM
F
P W
T
S
= M
P RF
P E
M
TT
R
S
0 X
I
P
XX
P
L
F
P
R
TT
P
I
P
U
I
M
P
I0 FT
W
F
P
TX
6
P W
T X
Signal 1
PRF
IF
PM
T
H
T X
E U ML S
O ED U
M LO A DT UO LR A T O
I
IT C
z
F
IT S
C W
H
IT C
T X
T X
H
TO R
IF = 6 0 0 M H z
ODULATOR
R F S W ITC H
z
RF SWITCH
TX
6. Proposed Fuzzy Controller
Fuzzy algorithm used in some of the literature above had two inputs and one output, some
even TX
used only one input and one output. Where as, for the process used one inference system,
that is Mamdani or Sugeno models. The proposed Fuzzy controller has two modules fuzzy,
those are fuzzy A module using mamdani model and fuzzy B module using Sugeno models.
Each module has two inputs and one output. Fuzzy rules in the fuzzy A module are based on
TX
the knowledge
of an expert and the experimental results. Meanwhile, fuzzy B module is
developed to read the output response of AGC and to maintain the module A output. The
detailed design will be explained below.
R F S W IT C H
Power Input
AGC
Gain
output
Fuzzy A
Maximum
Range
Fuuzy B
Power Output
AGC
Figure 9. Fuzzy Logic for AGC HM992LP5E
68
VCTRL
AGC
Eko Joni Pristianto, et al.
AGC HMC992LP5E gain control has two fuzzy modules: fuzzy A and fuzzy B module.
Fuzzy A module uses AGC Power input and maximum target range reading as the input and
generate output gain value. The input of fuzzy B module are fuzzy A output and AGC power
output signal. The block diagram of fuzzy logic for AGC is shown on Figure 9 below.
The first input of Fuzzy A is the maximum target range, which will be divided into 4 fuzzy
sets by triangle function as shown on Figure 10(a). The second input is AGC input amplitude
which is divided into 5 fuzzy sets as shown on Figure 10(b). Then, the output of Fuzzy A will
be divided into 4 fuzzy sets as shown on Figure 10(c). The Fuzzy A rules is shown on the
following Table 3. It can be shown on Table 3, there will be 20 fuzzy rules and the
defuzzyfication process of Fuzzy A module uses Mean of Maximum method.
(a)
(b)
(c)
Figure 10. Membership function (a) maximum range, (b) Input amplitude, (c) output gain
Maximum
radar target
range
Near
Moderate
Far
Very Far
Table 3. Fuzzy rules for Fuzzy A module
Power Input AGC
Very Small
Small
Moderate
Big
VeryBig
SR
SR
SR
SR
R
R
S
S
S
S
S
S
S
T
T
T
T
T
T
T
69
Automatic Gain Fuzzy Logic Controller for Pulse Radar Receiver System
Fuzzy A Output:
SR = very low output gain
R = low output gain
S = moderate output gain
T = high output gain
Fuzzy B uses Sugeno inference fuzzy model. The first input value of Fuzzy B is output gain
and the second input is AGC output amplitude. Each of inputs are divided into 4 and 5 fuzzy
sets with triangle function. Membership function of first and second input of Fuzzy B is shown
on Figure 11(a) and Figure 11(b).
(a)
(b)
(c)
Figure 11. (a) Membership function of output gain, (b) output signal level, (c) VCTRL
Output membership function of Fuzzy B module is VCTRL. That is divided into 7 fuzzy
sets as shown on Figure 11(c). Because Fuzzy B module uses Sugeno inference fuzzy model,
the fuzzy set output is a singleton function. Further rules for Fuzzy B module will be
constructed based on Table 4.
Gain output
Very low
Low
Moderate
High
Table 4. Fuzzy rules for Fuzzy B module
Output signal level
Very low
Low
Moderate
V2
V2.5
V3
V2
V2.5
V3
V1
V1.5
V2.5
V1
V1.5
V2.5
70
High
V3.5
V3.5
V3.5
V3
Very High
V4
V4
V4
V4
Eko Joni Pristianto, et al.
By defining fuzzy rules according to Table 4, there will be 20 fuzzy rules on Fuzzy B
module and the defuzzyfication process of Fuzzy B uses Weighted Average method.
7. The Implementation
The hardwares of AGC consist of several circuit blocks i.e. power supply, RF power
detector, PWM to analog converter, two-way splitter 600 MHz, AGC HMC992LP5E,
microcontroller, and LCD 20x4. Figure 12 shows the implementation of schematic circuit for
AGC HMC992LP5E gain control system with fuzzy logic. Power supply circuit is a 5V DC
regulator with 2A current using integrated circuit LM2576. RF power detector, AD8313, is
connected to input and output amplitude sensor. This circuit converts input and output into DC
signal.
Input
Splitter
AGC HMC992LP5E
Output
Splitter
Output Port
(to Spectrum)
Input Port
(to SigGen)
RF Power
Detector (output)
RF Power
Detector (input)
LCD 20x4
Volt meter
(VCTRL AGC)
USB Cable
(to PC)
Regulator
5V DC
Microcontroller
DC Input
9-15 V
PWM
to Analog
Figure 12. Implementation of AGC HMC992LP5E gain control system with fuzzy logic
RF power detector input and output will be connected to analog pin 0 (PA0) and analog pin
1 (PA1) of microcontroller. PWM-to-analog converter circuit uses LTC2645 and is used to
convert PWM value on pin PB10 of microcontroller into analog value with conversion time
about 8 us [13]. 600 MHz splitter is used to split RF signal with -3 dB attenuation on both
input and output side [14]. AGC HMC992LP5e will be controlled by fuzzy logic system. This
implementation used ARM 32 bit STM32F401RE microcontroller with LCD 20x4 circuit to
monitor sensor value and fuzzy calculation, which are connected to pin B9, B8, C9, C6, and C5
of microcontroller. Microcontroller will also be connected to computer through serial RS 232
channel.
8. Result and Analysis
Fuzzy A and Fuzzy B testing used 32 data samples for each maximum range value. Figure
13 shows relations between Fuzzy A output, from MATLAB and microcontroller calculation,
and input signal change. Figure 13(a) is a comparison between Fuzzy A output as a result of
microcontroller and MATLAB calculation for maximum target range of 20 NM. In this figure,
plot of microcontroller calculation coincide with plot of MATLAB calculation. From these
data can be concluded that average error on range of 24 NM is 2.39%.
Figure 13(b) shows result of Fuzzy A module testing on maximum target range which is
192 NM. On this range, the change Fuzzy A output is about -45 dBm to -15 dBm and the
generated Fuzzy A output is -12 dBm to -10 dBm. The significant change of output compared
to 24 NM range happens at input signal value -55 dBm to -45 dBm. On range of 24 NM, Fuzzy
A output is -60 dBm to -52 dBm. For 192 NM range, Fuzzy A output is -40 dBm to -35 dBm.
The average error at this range is 4.37%.
71
20,00
dBm)
Automatic Gain Fuzzy Logic Controller for Pulse Radar Receiver System
4,00
0,00
4,50
-10,00
-5,00
3,50
4,00
-10,00
3,00
3,50
Output Fuzzy B (Volt)
Output Fuzzy A (dBm)
Output Fuzzy A (dBm)
-20,00
-30,00
-40,00
-50,00
Fuzzy A_24 NM_uC
Fuzzy A_24 NM_Matlab
-60,00
Output Fuzzy B (Volt)
0,00
-15,00
2,50
-20,00
2,00
-25,00
1,50
-30,00
1,00
-35,00
-70,00
-60,00
-50,00
-40,00
-30,00
-20,00
-10,00
0,00
10,00
-45,00
0,00
-60
-60,00
-40
-40,00
-30
-30,00
-20
-20,00
-10
-10,00
0
0,00
10
10,00
-60
(b)
Figure 13. The Relations between Fuzzy A output and input signal
Figure 14 shows a correlation between Fuzzy B output from microcontroller and MATLAB
calculation and input signal change. Figure 14(a) shows a comparison of Fuzzy B output from
microcontroller calculation with MATLAB calculation at 24 NM range. There are some
difference between the two calculations. On the third data when the input is about -52 dBm. At
this point, MATLAB calculation generates Fuzzy B output of -25 Volts. Whereas,
microcontroller calculation generates -2.8 Volt value. At this range, the deviation often
happens at input signal about -10 dBm to 0 dBm and above. Average error at range of 24 NM
is 4.23%.
0,00
4,50
-5,00
4,00
-10,00
3,50
Fuzzy A_24 NM_uC
Fuzzy A_24 NM_Matlab
-10,00
0,00
10,00
Output Fuzzy B (Volt)
-15,00
Output Fuzzy A (dBm)
Output Fuzzy B (Volt)
3,00
2,50
-20,00
2,00
1,50
-25,00
-30,00
1,00
-35,00
0,50
-40,00
Fuzzy
B_24
NM_uC
Fuzzy
A_192
NM_uC
Fuzzy
B_24
NM_Matlab
Fuzzy
A_192
NM_Matlab
0,00 -45,00
-60
-60,00
-50,00
3,00
2,50
2,00
1,50
1,00
Fuzzy B_192 NM_uC
Fuzzy B_192 NM_Matlab
0,50
0,00
-50
-40,00
-40-30,00 -30-20,00 -20 -10,00 -10 0,00
DayaDaya
SinyalSinyal
Input Input
(dBm(dBm)
)
0
10,00 10
-60
-50
-40
-30
-20
Daya Sinyal Input (dBm )
-10
0
10
(a)
(b)
Figure 14. The Correlation between Fuzzy B output and input signal
Figure 14(b) shows Fuzzy B module testing on maximum target range that is 192 NM.
There are 4 points with significant deviation which are at input signal of -39, -32, -25, and 3
dBm. The average error at 192 NM is 3.96%. The average error of Fuzzy B module is greater
than the average error for Fuzzy A module because the second input value of Fuzzy B has a
feedback signal from AGC fuzzy system. This change will directly affect the calculation on
Fuzzy B module which results in the change of the generated output value. There are several
possible factors i.e. RF power detector output and the accuracy of ADC microcontroller.
The generated signal from pulse radar target simulator consist of 5 signals. For easy
observation, first the input signal is modulated by RF detector diode. By this process, a pure
pulse signal will be obtained without carrier signal and its negative part will be cut. The
generated signal from this process is shown on Figure 15(a). Signal 1 up to 5 has same width of
2 ms with the distance between signal is 10 ms. Amplitude of each signal is distinguished to
test the response of AGC system. The amplitude of each signal are:
Signal 1
= 80 mV (-8,93 dBm),
72
1,50
0,50
DayaSinyal
SinyalInput
Input (dBm)
(dBm)
Daya
(a)
3,50
2,00
0,00
-50
-50,00
Daya Sinyal Input (dBm)
4,00
2,50
1,00
Fuzzy B_24 NM_uC
Fuzzy A_192 NM_uC
Fuzzy
B_24NM_Matlab
NM_Matlab
Fuzzy
A_192
0,50
-40,00
3,00
-50
Eko Joni Pristianto, et al.
Signal 2
Signal 3 and 4
Signal 5
= 100 mV (-6,99 dBm),
= 20 mV (-20,99 dBm),
= 50 mV (-13,01 dBm)
The testing of classic AGC system is illustrated on Figure 15(b). According to the
characteristics of classic AGC system, all input signal amplitudes within dynamic range will be
fixed to a certain point set. In this system, the magnitude of signal 1 to signal 5 have same
amplitude that is 150 mV (-3.47 dBm). On pulse radar receiver system, this system can be used
for pulse radar that only calculates radar target range. It will be less effective for pulse radar
which calculates target range and the intensity of detected target amplitude.
(c)
Figure 15. (a) Input signal, (b) the Output signal of classic AGC system,
(c) AGC Fuzzy system output signal on maximum target range at 192 NM
Figure 15(c) shows AGC fuzzy system output signal on maximum target range at 192 NM.
If it is compared to the classic AGC system, there are output signal differences:
Signal 1 has initially value 80 mV (-8.92 dBm) and increases to 150 mV (-3.47 dBm).
Signal 2 has initially value 100 mV (-6.99 dBm) and increases to 140 mV (-4.07 dBm).
73
Automatic Gain Fuzzy Logic Controller for Pulse Radar Receiver System
Signal 3 and 4 has initially value 20 mV (-20.99 dBm) and both increase to 150 mV (-3.47
dBm).
To prove this fuzzy system has worked well, it can be analyzed on signal 3. When input
signal is at -20.99 dBm and maximum target rang is 192 NM, Fuzzy B output will generate
VCTRL value about 2.5 V. According to AGC gain characteristic, this VCTRL value will
amplify AGC to 13 dB. So, the signal 1 which has amplitude -20.99 dBm, when passing
through AGC fuzzy will have amplified -7.99 dB. This can bee seen in the measured signal on
the osiloscope.
Output signals on classic AGC system will have the same amplitude when the system is
given varied amplitude input signal. Meanwhile, output signal amplitude of AGC fuzzy
system is calculated according to the fuzzy algorithm. As the result, output signal amplitude is
dynamic. The most important of these two fuzzy systems – classic AGC and Fuzzy AGC – is
AGC output must have same width and distance between signals with its input. This is
important to prevent target range calculation error in pulse radar system application. This is
already done at the testing, where the width of output does not change, which is 2 ms and the
distance between signals is 10 ms.
9. Conclusion
In this paper, AGC HMC992LP5E gain controller with fuzzy logic has been designed and
implemented successfully on microcontroller with input variables are AGC input-output signal
and maximum radar target range. The average error of Fuzzy A module is 3.51% and Fuzzy B
is 4.47%. This system has been tested using pulse radar target simulator. AGC Fuzzy system
has only changed the input signal amplitude and kept constant the width and distance between
signals. This system is more effective to be applied on pulse radar systems which calculates
target range and intensity of detected target’s amplitude. Thus, the target captured by receiver
will be more easily identified.
10. References
[1]. Skolnik, M. I.,“Introduction to Radar Systems”, 3rd ed, 1-5, McGraw-Hill Book Company
Inc., 2001.
[2]. ”Automatic Gain Control Methods”,
http://www.radartutorial.eu/09.receivers/rx08.en.html, Accessed at 1 May 2015.
[3]. “Datasheet IC HMC992LP5E”, v03.1013, Hittite Microwave Coorporation,
https://www.hittite.com/content/documents/data_sheet/.
[4]. Design of air surveillance radar, Radar Consortium (2014), 2nd Report Progress Sinas
Research, PPET-LIPI, Bandung.
[5]. http://www.intechopen.com/books/aeronautics-and g stronautics/the-assessment-methodfor-multi-azimuth-and-multi-frequency dynamic-integrated-stealth-performance-o, 25
Mei 2015.
[6]. Kim, J.W., Zhang, L., Seo, J.Y., Cho, H.C., Seo, H.I., Cho, T.H., dan Jung, J.D.,
“Improved Automatic Gain Control Circuit Using Fuzzy Logic”, Fuzzy Systems and
Knowledge Discovery, pp. 159-168, 2006.
[7]. Barajas, J.R., Assad, G.D., dan Soto, R., “A fuzzy logic based AGC algorithm for a radio
communication system”, The Ninth IEEE International Conference on Fuzzy Systems, vol
2, pp 977-980, 2000.
[8]. Murat, Y., Goktas, H.H., dan Celebi, F.V., “Design and implementation of fuzzy logic
based automatic gain controller for (EDFAs)”, International Journal for Light and
Electron Optics, vol 125, issue 18, pp 5450–5453, 2014.
[9]. Delgado, L.G., Redrovan, D.V., Bykbaev, V.R., Delgado, N.G., dan Panzner, T., “Fuzzy
Controller for Automatic Microphone Gain Control in an Autonomous Support System
for Elderly”, The 2nd International Workshop on Service Science for e-Health (SSH), pp
77-81, 2014.
[10]. Suyanto, Soft Computing, 27-56, Informatika, Bandung, 2008
74
Eko Joni Pristianto, et al.
[11]. Zadeh, L. A., “Fuzzy Sets”, Journal Information and Control, 8, 338-353, 1965.
[12]. Datasheet
IC
AD8313,
0.1
2.5
GHz,
Analog
Device,
http://www.analog.com/en/products/rf-microwave/rf-power-detectors/non-rmsresponding-detector/ad8313.html
[13]. Demo Manual DC2197A, Linear Technology, http://cds.linear.com/docs/en/demo-boardmanual/DC2197AF.PDF, 5 Mei 2015.
[14]. Marbun, A. J., Design of Chebyshev Power Combiner 2:1 at 2400 Mhz, Thesis,
Universitas Indonesia, 2008.
Eko Joni Pristianto, is a researcher at Research Center for Electronics and
Telecommunications LIPI, Indonesia. He was graduated from Universitas
Jember (UNEJ) in 2008. He obtained his master degree from Electrical
Engineering program, Bandung Institute of Technology (ITB), in 2015. His
research interests include radar system, underwater communication system,
automation system and control system.
Pranoto Hidaya Rusmin was born in Magelang, Indonesia in 1972. He
received B.Eng., M.Eng., and Doctor degrees in electrical engineering from
Institut Teknologi Bandung (ITB), Indonesia, in 1996, 1999, 2009,
respectively. Since 1998, He is a Lecturer at School of Electrical Engineering
and Informatics ITB, Indonesia. His research interest is Internet Congestion
Control.
75