Uploaded by anantia prakasa

The Performance of Supervised and Unsupe

advertisement
SEMINAR RADAR NASIONAL III
2009
Prosiding
Bandung
30 April 2009
Savoy Homann Bidakara Hotel
Penyelenggara:
ppet
Pusat Penelitian Elektronika dan Telekomunikasi
Lembaga Ilmu Pengetahuan Indonesia (PPET - LIPI)
LIPI
dan
Sekolah Teknik Elektro dan Informatika (STEI - ITB)
bekerjasama dengan
International Research Centre for Telecommunications
and Radar (IRCTR) Delft University of Technology
(TU Delft) The Netherlands
Sponsor:
i
Prosiding
Seminar Radar Nasional III 2009
ISSN : 1979 - 2921
Hak cipta © 2009 oleh Pusat Penelitian Elektronika dan Telekomunikasi – LIPI
Hak cipta dilindungi undang-undang. Dilarang menyalin, memproduksi dalam segala
bentuk, termasuk mem-fotocopy, merekam, atau menyimpan informasi, sebagian atau
seluruh isi dari buku ini tanpa ijin tertulis dari penerbit.
Prosiding Seminar Radar Nasional / [editor by] Mashury Wahab, Asep Yudi Hercuadi A.A.
Lestari, A.B. Suksmono, , Yuyu Wahyu, Pamungkas Daud.
vi + pp.; 21,0 x 29,7 cm
ISBN : 1979 - 2921
Radio Detecting and Ranging (Radar)
Technical editing by Yudi Yuliyus Maulana, Dadin Mahmudin, Deni Permana K,Yadi
Radiansah, Sulistyaningsih, Folin Oktaviani.
Cover design by Yadi Radiansah.
Diterbitkan oleh :
Pusat Penelitian Elektronika dan Telekomunikasi (PPET)
Lembaga Ilmu Pengetahuan Indonesia (LIPI)
Kampus LIPI Jl. Sangkuriang, Bandung
Telp. (022) 2504661 Fax. (022) 2504659
Website : www.ppet.lipi.go.id
ii
Pelindung
Deputi Bidang Ilmu Pengetahuan Teknik LIPI
Ketua Umum
Mashury Wahab
Panitia Pengarah
Lilik Hendradjaja, Dephan
Adang Suwandi, ITB
Syahrul Aiman, LIPI
Tatang A. Taufik, BPPT
Hiskia Sirait, LIPI
Andriyan B. Suksmono, ITB
A. Andaya Lestari, IRCTR-IB
Endon Bharata, IRCTR-IB
Nana Rachmana, ITB
Yuyu Wahyu, LIPI
Syamsu Ismail, LIPI
Rustini S. Kayatmo, LIPI
Eko Tjipto Rahardjo, UI
Adit Kurniawan, ITB
Panitia Pelaksana
Iskandar, ITB
Andi Kirana, RCS
Gunawan Handayani, ITB
Pamungkas Daud, LIPI
Folin Oktaviani, LIPI
Sulistyaningsih, LIPI
Yudi Yulius Maulana, LIPI
Dadin Mahmudin, LIPI
Deni Permana K., LIPI
Sri Hardiati, LIPI
Iqbal Syamsu, LIPI
Asep Yudi Hercuadi, LIPI
Yadi Radiansah, LIPI
Zaenul Arifin, LIPI
Endang Ridwan, LIPI
Lisdiani, LIPI
Poppy Sumarni, LIPI
Noorfiya Umniyati, LIPI
Dicky Desmunandar, LIPI
iii
Daftar Makalah
1.
Tinjauan Pemanfaatan Teknologi Seluler GSM Diaplikasikan Sebagai
Pasif Radar .............................................................................................................
Febrian Wijoseno, Adit Kurniawan, Arwin D.W.Sumantri
2.
Analisa FDOA-RADAR Sekunder Terhadap Gangguan Random Noise
Wahyu Widada dan Sri Kliwati …………………………………………………………..
3.
Sistem Identifikasi Pesawat Menggunakan Kecepatan dan Radar Cross
Section Pesawat Berbasis Jaringan Syaraf Tiruan Backpropagation …………
Maman Darusman, Arwin Datumaya Wahyudi Sumari, Aciek Ida Wuryandari
4.
5.
The Performance Of Supervised And Unsupervised Neural Networks In
Performing Aircraft Identification Tasks …………………………………………….
Arwin Datumaya Wahyudi S, Aciek Ida Wuryandari, Maman Darusman, Nur
Ichsan Utama
Minimalisasi Sinyal Harmonik Radar Indra I dengan IF Lumped Element
Filter ........................................................................................................................
Liarto, A.A Lestari, Sri Hardiati
1-5
6-10
11-15
16-22
23-26
6.
First Results Of The Signal Processing Of INDERA ……………………………..
W. Sediono, A. A. Lestari
7.
Teknik Pengukuran Pola Radiasi Transduser Dwi-Fungsi Akustik
Bawah Air …………………………………………………………………………………
Syamsu Ismail
30-32
Analisis Penyesuai Impedansi NTL Menggunakan Metode Ekspansi
Fourier .....................................................................................................................
Rudy Yuwono, Achmad Setiawan, D.J. Djoko H.S
33-38
Antena Array Electronic Switch Beam Untuk Pengerahan Beam Antena
Pada Sistem Radar .................................................................................................
Yoko Wasis, Bambang Edhie Sahputra, Iswahyudi, Yogi Koswara, A.A Pramudita
39-41
8.
9.
27-29
10.
Antena UWB Bentuk T Untuk Aplikasi SFCW-GPR 100-1000 Mhz ...................
A.Adya Pramudita, A. Kurniawan, A. Bayu Suksmono, A.Andaya Lestari
42-46
11.
Sistem Trigger Pada Radar Maritim Indera ..........................................................
Oktanto Dedi Winarko, A. Andaya Lestari
47-51
12.
Penelitian-Penelitian Radar Dan Pendukungnya di Lipi .....................................
Masbah R.T. Siregar
52-56
13.
Karakterisasi Penggunaan Garis Kurva Pada Lengan Seri 3 dB Hybrid
Coupler MicrostripPita Lebar ................................................................................
Y.K. Ningsih, F.Y. Zulkifli, E.T. Rahardjo, A.A. Lestari
57-62
Optimasi Pemodelan Arima Dengan Efek Deteksi Outlier Pada Data
Curah Hujan Di Surabaya ......................................................................................
Achmad Mauludiyanto, Gamantyo Hendrantoro, Mauridhi Hery P, Suhartono
63-67
Desain Dan Simulasi Tranceiver Stepped Frequency ContinuousWave
Ground Penetrating Radar (SFCW GPR) 700 – 1400 Mhz ..................................
Tommi Hariyadi, Endon Bharata, Andriyan Bayu Suksmono
68-74
14.
15.
vi
16.
Sistem Antena Radar VHF Lapan .........................................................................
Peberlin Sitompul, Aries Kurniawan, M. Sjarifudin, MarioBatubara, Harry Bangkit,
Timbul Manik, J.R Roettger
17.
Pembuatan Modul Receiver Untuk Sistem Perangkat Pemancar
Jamming …………………………………………………………………………………..
Elan Djaelani
79-84
Pemanfaatan Sistem Pakar dalam Perancangan Sistem Analisa Masalah
Dan Penentu Tindakan Pemeliharaan Radar .......................................................
Edith Nurhidayat Kurniawan S. , Aciek Ida Wuryandari , Arwin D.W. Sumari
85-88
18.
75-78
19.
Penggunaan Radar Bagi Kepentingan Pertahanan Udara .................................
Suparman D, MM
20.
Kerjasama DEPHUT Dan Lembaga International dalam Penggunaan Radar
Untuk Mendukung Pengelolaan Hutan yang Lestari ……………………………...
Iwan Setiawan Priyambudi Santoso
95-96
Perbandingan Performansi Sistem Identifikasi Pesawat Menggunakan
Jaringan Syaraf Tiruan Mode Adaptive Resonance Theory 1 Dan 2 .................
Nur Ichsan Utama, Aciek Ida Wuryandari, Arwin D. W. Sumari
97-101
Ilmu Pengetahuan Rekayasa dan Teknologi Dan Seni (ILPERTEKS)
Untuk Pengembangan Radar Pengawas Pantai .................................................
,
Elan Djaelani Prof. Dr. Rohani J Widodo, Ridodi Anantaprama, Iwan Setiawan
102-108
21.
22.
89-94
23.
Usulan Pemakaian Radar Langit Untuk Daerah Khusus Atau Rawan ..............
Hari Satriyo Basuki
109-113
24.
Implementasi Peta Dinamis Pada Radar INDERA ...............................................
Deni Yulian, W. Sediono, A. Andaya Lestari
114-118
25.
Perancangan dan Realisasi Antena Rolled Dipole Untuk Keperluan
Ground Penetrating (GPR) Dengan Menggunakan Metoda Finite
Different Time Domain (FDTD) ..............................................................................
Yudi Yuliyus Maulana, Yuyu Wahyu, Folin Oktafiani dan AA Lestari
119-124
26.
Antena Ground Penetrating Radar Adaptif Terhadap Multi Pulsa .....................
Folin Oktafiani, Yuyu Wahyu, Yudi Yuliyus, A.A Lestari
27.
Estimasi Electromagnetic Interference (EMI) Dalam Sistem Antena Patch
Array Untuk Radar .................................................................................................
Sri Hardiati, Sulistyaningsih
129-132
Antena Dipole Dengan Pembebanan Resistif dan Layer Dielektrik Untuk
Ground Penetrating Radar (GPR) .........................................................................
Y.Wahyu, A.Kurniawan, Sugihartono, A.S Ahmad, A A Lestari
133-137
28.
125-128
29.
Kajian Mengenai Radar Clutter Dan Pengaruhnya Pada Unjuk Kerja Radar ...
Mashury Wahab dan Sulistyaningsih
138-142
30.
Pembangkit Chirp Untuk Radar FMCW Menggunakan DDS ..............................
Purwoko Adhi
143-146
vii
The Performance of Supervised and Unsupervised
Neural Networks in Performing Aircraft Identification Tasks
Arwin Datumaya Wahyudi Sumari 1) Aciek Ida Wuryandari 2)
Maman Darusman 2) Nur Ichsan Utama 2)
1) Departemen Elektronika, Akademi Angkatan Udara
Jl. Laksda Adisutjipto, Yogyakarta 55002 – INDONESIA
Telp. 0274 486922 ext 6101 Fax. 0274 488918 Email: arwin91aau@yahoo.co.id
2) Sekolah Teknik Elektro dan Informatika - ITB,
Kampus ITB Labtek VIII Lantai 2, Jl. Ganesa 10, Bandung 40132 – INDONESIA
Telp. 022 2502260 Fax. 022 2534222 Email: aciek@lskk.ee.itb.ac.id,
maman_darusman@yahoo.com, nur_ichsan@ymail.com
ABSTRACT
This paper is a report on our research progress in the area of aircraft identification by utilizing neural networks
and information fusion. In this paper we address the performance comparison of supervised and unsupervised
neural networks in aircraft identification tasks in a generic system called Neural Network-based Aircraft
Identification System (NN-AIS). We select Adaptive Resonance Theory (ART) for the unsupervised neural
network and Back Propagation Network (BPN) for the supervised one. As for previous research, we use two
kinds of input namely aircraft Radar Cross Section (RCS) and average speed. Their performance will be
validated by using already-learnt and never-learnt patterns.
Keywords: ART, BPN, aircraft identification, RCS, speed
In this paper we address the utilization of two
kinds of neural network architecture for performing
aircraft identification tasks namely supervised and
unsupervised. We use the two types of input for
training and validating the networks and measure their
performance. In general our proposed system is called
as Neural Network-based Aircraft Identification
System (NN-AIS).
The structure of the paper is as follows. Section 1
cover the background of the paper and it will be
followed by Section 2 which covers a short
introduction to neural networks as well as related
matters to aircraft RCS and speed. Section 3 presents
the NN-AIS design as well as the NN training. Section
4 delivers the system validation. The paper is
summarized in Section 5 with some concluding
remarks.
1. INTRODUCTION
Aircraft identification task is a critical matter to
recognize the identity of an aircraft that is entering a
monitored air space. The sooner the observed aircraft
is identified, the faster the authorized authority can
make a decision. In normal flight procedure, all
aircraft flight plans must be reported to the authorized
authority to be recorded. The records will be used to
monitor every single aircraft movement in the
monitored air space.
The reported flight plans ease the authority to track
and identify a certain aircraft that is displayed on
monitor room’s displays. A problem is arisen when
the authority cannot identify a certain aircraft that is
detected by radar system. There are two possibilities
when an aircraft cannot be identified. First, there is a
possibility the aircraft transponder for answering the
interrogation signal from ground station is not
working properly or failed. Second, there is an
intention to turn-off the transponder in order to hide
the aircraft identity.
For the second reason, we can conclude that the
aircraft must be in undercover missions and can be a
threat to our country’s sovereignty. Because of it, the
authority must have a way to identify the unidentified
aircraft before something harmful occurs in the future.
One of ways in identifying aircraft is by using its
Radar Cross Section (RCS) value and combining it
with its speed. On the other hand, one approach that
has been known well for object recognition is neural
networks.
2. A SHORT INTRODUCTION TO NEURAL
NETWORK AND AIRCRAFT RCS
In this section, we will present a short introduction
to neural network, its learning paradigm taxonomy,
and a brief explanation regarding two types of
learning paradigm we select for our research. We also
deliver a very brief explanation regarding aircraft RCS
and speed.
2.1. Neural Network
The most basic constructing element of a human
nervous system is a neuron which is called as
“processing unit” as presented in Figure 1. According
16
to Shepherd and Koch (1990) in Haykin (1994),
human brain has more than 10 billion neurons and 60
trillion synapses or connections between neurons.
Even though it is relatively slower than computer
systems that are made up from nano-technology
silicon gates, it can do highly complex, nonlinear, and
parallel tasks such as pattern recognition and
perception, faster and very much better than the best
computing system that human ever created.
In NN model, neuron takes a set of inputs, xm ,
along with a set of connection or link or synaptic
which are characterized by weights, w km . The
summing junction, ∑, sums up the input signals that
are amplified by the connection weights. The
activation function, ϕ ( . ) , limits the net outputs in
allowable values. The architecture of the NN model is
depicted in Figure 2.
The general mathematical equations for neural
information processing are given in Equation 1 for
inputs summing process to obtain v k
vk = ∑ wkj x j
m
(1)
j =0
and Equation 2 for producing the NN output, yk .
yk = ϕ ( v k )
(2)
Figure 1: Neuron or nerve cell.
2.2. Neural Network Learning Model Taxonomy
Because the neural networks are good for
recognition tasks, some earlier researchers such as
McCulloch-Pitts, Grossberg, Minsky, etc., tried to
model the nervous processing unit so its mechanism
can be emulated in computing systems. From this
perspective, we define Artificial Neural network
(ANN) or usually called as just Neural Network (NN)
as an emulation of human nervous system when
performing information processing. Its characteristics
are displayed on the ability to obtain new knowledge
after a successful learning process and store it in the
information storage which is its synaptic weights.
In more detail, NN is generalization of
mathematical models of human cognition based on the
assumptions that (Fausset, 1994):
• information processing occurs at many simple
elements called neuron,
• signals are passed between neurons over
connection links,
• each connection link has an associated
connection weight which multiplies the signals
transmitted,
• each neuron applies an activation function
which is usually non linier, to its net input to
determine its output signals.
According to Haykin’s (1994) taxonomy, there
are three NN learning models.
• Supervised. The essential of this paradigm is
the availability of an external supervisor, so
there will be an input-output relation in order to
find the most minimum disagreement between
the NN outputs with the examples given by the
supervisor.
• Unsupervised or Self-Organized. In this
learning paradigm, there is no external teacher
or examples to be learnt by the NNs. So, the
NNs will perform a competitive learning rule
where the winning neuron is entitled to keep
the input in its memory.
• Reinforcement Learning. This is the on-line
learning of an input-output mapping through a
process of trial and error designed to maximize
a scalar performance index called as
reinforcement signal.
2.3. Supervised Neural
Propagation Network
Network
–
Back
The BPN was developed to cope with the
limitations of single-layer NN. The BPN actually is a
feedforward NN that is trained by backpropagation
which means the signals is propagated in reverse
direction. The primary aim of NN training is to train
the NN to achieve a balance between the ability to
respond correctly to the input patterns that are used for
training or memorization, and the ability to give
reasonable responses to input that is similar that used
in training or generalization.
In training the NN with backpropagation
mechanism, there will be three steps, namely:
• feedforwarding the input training patterns to the
NN input layer,
Figure 2: A mathematical model of a neuron.
17
•
Figure 4: ART1-NN architecture (Sumari et.al, 2008a).
calculating the NN outputs and backpropagating
the associated error,
• adjusting the connection weights to minimize
the error.
After passing the training session, the NN uses the
final connection weights when recognizing the input
patterns given to it.
As commonly in a multilayer NN, BPN has three
layers namely input layer, hidden layer, and output
layer. The number of input and output layers is
depended on the input pattern and the output’s target.
The number of hidden layer is depended on particular
applications, but commonly one hidden layer is
sufficient for many applications. The architecture of
BPN is depicted in Figure 3.
ART-NN, as presented in Figure4, is designed to
facilitate degree controlling of pattern similarity that is
placed at the same cluster and overcomes stabilityplasticity problem that is faced by other NN. The
ART1-NN is also designed to group (cluster) binary
input vectors. It has two layers, F1 layer is divided
into two sublayers namely F1(a) as input part, and
F1(b) as interface part, and layer F2 (cluster) along
with reset unit that is used to control degree of
patterns similarity that is put down at the same unit
cluster. F1 and F2 layers are connected by two groups
of weight paths, bottom-up weight and top-down
weight. To control learning process, some
complement units are also entangled at this NN.
For performing pattern matching, ART is provided
with a parameter called vigilance parameter, ρ with a
value in range 0 < ρ ≤ 1 . Higher values of ρ are
applied for training session, while lower values are
applied for operating session.
ART1-NN architecture consists of two parts. The
architecture of the ART1-NN is depicted in Figure 4.
• Computation Units. It consists of F1 layer
(input part and interface part), F2 layer, and
reset unit. The F2 layer is also called as
competitive layer.
• Complementary Units. This unit provides a
mechanism so that the computation that is
carried out by ART1 algorithm can be done by
using NN principles. These units are also called
as gain control units. For the ART1-NN
algorithm in detail, refer to (Skapura, 1991).
Figure 3: The architecture of the BPN model (Priyanto
et.al, 2008).
1) Training Algorithm. Refer to (Fausset,
1994) for detailed BPN training algorithm.
2) Activation Function. There are four
common NN activation functions namely:
• identity function which produces binary
output (0 or 1),
• binary step function with threshold θ ,
• binary sigmoid (logistics sigmoid),
• bipolar sigmoid,
• hyperbolic tangent.
The sigmoid function and hyperbolic tangent are
the most common activation function for training NN
with backpropagation mechanism.
2.5. Aircraft Radar Cross Section and Speed
Radar Cross Section (RCS) is comparison between
power density that is reflected to transmitting source
and power density that is reflected by detected target
or object. Figure 5 shows example of aircraft RCS
that captured by Radar. Every aircraft or air object
has sharply differentiated RCS in accordance with
configuration elements that form RCS itself.
2.4. Unsupervised Neural Network – Adaptive
Resonance Theory
18
Figure 6: The concept of information fusion (Ahmad &
Sumari, 2008).
The information sources can be from as follows:
• observation data from distributed sensors,
• commands and data from operator or user,
• a priori data from an existing database.
Referring to (Hall, 2001) in (Ahmad & Sumari,
2008), for obtaining a comprehensive information in
decision level, we can select many technique options
such as Boolean operator methods (AND, OR) or
heuristics value such as M-of-N, maximum vote, or
weighted sum from hard decision and Bayes method,
Dempster-Shafer, and fuzzy variable for soft decision.
In this paper we use the Boolean operator for all
approaches.
Figure 5: Common aircraft RCS (Nopriansyah et.al, 2008).
Aircraft speed that is presented at Radar screen can
be obtained by using Doppler principle that is shown
in Equation 3. Figure 3 illustrates how to calculate
object speed by using Doppler principle.
fd =
2.v
λ
cos θ
(3)
which fd is Doppler shift, v is aircraft speed, λ is
wavelength, and θ is angle between direction of
incoming signal propagation with direction of antenna
movement.
In this paper we use RCS and speed data taken
from previous research done by Nopriansyah et.al,
(2008) as presented in Table 1.
3. A GENERIC MODEL OF NEURAL
NETWORK AIRCRAFT IDENTIFICATION
SYSTEM
The NN-AIS generic model is modified from
Sumari et. al (2008b) which consists of three
processing blocks namely Pre-processing Block,
Aircraft Identification Block, and Post-processing
Block.
Table 1: List of aircraft RCS and speed data
(Nopriansyah et.al, 2008).
No.
1.
2.
3.
4.
5.
6.
7.
8.
Aircraft Type
Bell 47G
F-16 Fighting Falcon
Hawk200
Su-30 Sukhoi
Cobra AH-1S
Cassa C-212
CN-235
A-310 Airbus
RCS
3
5
8
15
18
27
30
100
Speed
(km/hour)
168.532
1470
1,000.08
2,878.75
227.796
364.844
459.296
980
3.1. NN-AIS Diagram Block
2.6. A Brief Introduction to Information Fusion
In general, information fusion is a technique in
combining physical or non-physical information form
from diverse sources to become single comprehensive
information to be used as a basis for prediction or
estimation of a phenomenon. The prediction or
estimation is then used as the basis for performing
decisions or actions. Figure 6 illustrates the concept of
information fusion.
Figure 7: Generic NN-AIS architecture.
Because the system processes two different data,
so there will be two NNs within the system, one is for
processing aircraft RCS data and the other is for
processing aircraft speed data. The generic
architecture of NN-AIS is depicted in Figure 7.
19
In general, the Pre-processing Block prepares the
inputs in form of vector patterns to the two NNs. The
Aircraft Identification Block performs identification of
the received inputs to the knowledge stored in the
NNs. The Post-processing Block fuses the output
resulted from Aircraft Identification Block, converts
and displays the estimated identity of the received
inputs.
The peculiar feature of an NN is if it already
learned the received input, it will produce an exact
result. But if it has never learnt the received input, it
will try to find the best match or estimated result.
3.2. Training in Supervised NN-AIS (Darusman
et.al., 2009)
In the NN1 architecture we use 150 neurons in
hidden layer, while in the NN2 architecture we use 20
neurons. We take these numbers after carrying out
some observations to find the most appropriate
numbers. For training the two NNs, we did some
researches to find the most appropriate activation
functions to be utilized to the NN architectures. We
select the combination of logsig and purelin activation
functions for hidden layer and output layer.
The BPN needs some time to train its structure in
order to learn the vector patterns given to it by
minimizing the difference error between the net’s
outputs with the target’s values. In order to train the
supervised BPN, we created vector patterns as
presented in Table 2 and Table 3. The results of the
NNs training are depicted in Figure 8 and Figure 9.
Figure 8: NN1 knowledge for aircraft speed data.
Table 2: Aircraft speed as inputs to NN1 and
the NN1 learning’s targets.
No.
1.
2.
3.
4.
5.
6.
7.
8.
Aircraft Speed
168.532
1470
1,000.08
2,878.75
227.796
364.844
459.296
980
Target
00000001
00000010
00000100
00001000
00010000
00100000
01000000
10000000
Figure 9: NN2 knowledge for aircraft RCS data.
3.3. Training in Unsupervised NN-AIS (Utama
et.al, 2009)(Wuryandari et.al, 2009)
In the same as the supervised one, we also created
vector patterns for unsupervised ART but in different
perspective as presented in Table 4 and Table 5.
Table 4: Aircraft RCS data and its associated pattern
Table 3: Aircraft speed as inputs to NN2 and
the NN2 learning’s targets.
No.
1.
2.
3.
4.
5.
6.
7.
8.
Aircraft RCS
3
5
8
15
18
27
30
100
No.
1.
2.
3.
4.
5.
6.
7.
8.
Target
00000001
00000010
00000100
00001000
00010000
00100000
01000000
10000000
20
Aircraft Type
Bell 47G
F-16 Fighting Falcon
Hawk 200
Su-30 Sukhoi
Cobra AH-1S
Cassa C-212
CN-235PTDI
A-310 Airbus
RCS
3
5
8
15
18
27
30
100
Vector Pattern
000000000011
000000000101
000000001000
000000001111
000000010010
000000011011
000000011110
000001100100
Table 5: Aircraft speed data and its associated pattern
Table 6: Validation data for supervised NNs.
No.
1.
2.
3.
4.
5.
6.
7.
8.
Aircraft Type
Bell 47G
F-16 Fighting Falcon
Hawk 200
Su-30 Sukhoi
Cobra AH-1S
Cassa C-212
CN-235 PT DI
A-310 Airbus
Average
speed
(km/h)
158.532
1,470
1000.08
2,878.75
227.796
364.844
459.296
980
Vector
Pattern
Input Data
000010011111
010110111110
001111101000
101100111111
000011100100
000101101101
000111001011
001111010100
Speed
RCS
1st
Detection
168
3
2nd
Dectection
169
3.2
3rd
Detection
167
2.7
Table 7: Validation result.
Output Types
Speed
RCS
Final result
The results of the unsupervised NNs training are
depicted in Figure 10 and Figure 11.
Aircraft
Number of
Type
Information
Bell 47G
3
Bell 47G
3
Bell 47G
Figure 12: The identification process carried out by the
supervised NNs.
Figure 10: Unsupervised NN1 knowledge in form of topdown and bottom-up weights for speed pattern.
4.2. Unsupervised NN-AIS
To validate the unsupervised NN-based system,
we select three aircrafts in random manner, namely
Cobra AH-1S and Bell 47G. In this validation we set
up the vigilance parameter ρ = 0.5. The results are
presented in Figure 13and Figure 14.
Figure 11: Unsupervised NN2 knowledge in form of topdown and bottom-up weights for RCS pattern.
Figure 13: Identification process result for Cobra AH-1S.
4. NN-AIS VALIDATION
In this section we present the NN-AIS validation
on the two types of NNs we already explained.
4.1. Supervised NN-AIS
Figure 14: Identification process result for Bell 47G.
To validate the unsupervised NN-based system, we
select three aircrafts in random manner, namely Bell
47G as presented in Table 6. In this validation we
modify the speed and RCS inputs to see if the system
works as it is designed. The results are presented
Table 7 and the identification process is depicted in
Figure 12.
4.3. Measuring the Performance
4.3.1.
Supervised NN-AIS
As we can see from the validation presented in
Table 5, the NNs in supervised NN-AIS tries to
21
recognize the detected aircraft’s patterns by
generalizing the knowledge of what they “see” and
what they have already learnt during training session.
The result is the system is able to produce the correct
aircraft estimation, namely Bell 47G helicopter.
4.3.2.
[4]
Unsupervised NN-AIS
Figure 14 and Figure 15 clearly present the
mechanism carried out by unsupervised NN-AIS in
identifying the detected aircrafts. The identification is
done directly by matching the detected aircrafts’
patterns with the knowledge they already memorized
during the training session. The result is the system is
able to produce the correct aircraft estimation, namely
Bell 47G and Cobra AH-1S helicopters.
[5]
[6]
[7]
5. CONCLUDING REMARKS
We have presented the utilization of supervised
BPN and unsupervised ART NNs and observe their
performances in identifying the identity of aircrafts in
NN-AIS framework. The two approaches result in
good estimations even though the vector patterns have
been modified. The supervised NNs use the
generalization capability to recognize the patterns
while the unsupervised ones use matching mechanism
in recognizing the patterns.
By noticing the results of this research, there is a
possibility that these approaches can be combined
with the real-life Radar system in order to increase its
identification tasks. The NNs-based system can give
significant advantage especially to identify harmful
unlisted detected aircrafts.
[8]
[9]
REFERENCES
[10]
[1] A.I. Wuryandari, A.D.W. Sumari, and
Nopriansyah, Aircraft Identification by Using
Combination of Neural Network and Information
Fusion, to be appeared in Jurnal Penelitian dan
Pengembangan Telekomunikasi (JURTEL), No.
2, Vol. 13, Desember 2008, ITTelkom, Bandung
[2] A.S. Ahmad, and A.D.W. Sumari, Multi-Agent
Information Inferencing Fusion in Integrated
Information System, Seri “Information Science
and Computing”, Sekolah Teknik Elektro dan
Informatika, Institut Teknologi Bandung,
Penerbit ITB, 2008, ISBN 978-979-1344-31-9.
[3] A.D.W. Sumari, et.al., Application of Adaptive
Resonance Theory 1 for Identification Friend,
Foe, or Neutral System, Proceedings of the 4th
International Conference Information &
Communication Technology and System 2008
(ICTS2008), Institut Teknologi 10 Nopember
[11]
[12]
22
Surabaya, Surabaya, 5 August 2008a, pp. 602609, ISSN 1858-1633.
A.D.W. Sumari, A.S. Ahmad, A.I. Wuryandari,
and Nopriansyah, Object Identification by Using
Combination of Neural Network and Information
Fusion, Proceedings of the 1st International
Graduate Conference on Engineering and
Science
2008
(IGCES2008),
Universiti
Teknologi Malaysia, Johor, Malaysia, D31, 2324 December 2008b, ISSN 1823-3287.
D.L. Hall and J. Llinas, Eds., Handbook of
Multisensor Data Fusion, USA: CRC Press LLC,
2001.
D.M. Skapura, Artificial Neural Networks:
Algorithms, Applications, and Programming,
Addison-Wesley, 1991.
D. Priyanto, A.D.W. Sumari, and E.P.T.
Wibowo, Design of Neural Network-based
Intelligent Classroom System: A Preliminary
Research, Proceedings of the 1st Makassar
International
Conference
on
Electrical
Engineering
and
Informatics
2008
(MICEEI2008), Universitas Hasanuddin, 13-14
Nopember 2008, Makassar, pp. 67-72, ISBN
978-979-18765-0-6.
L. Fausset, Fundamentals of Neural Networks:
Architectures, Algorithms, and Applications,
Prentince-Hall, USA, 1994.
M. Darusman, A.D.W. Sumari, and A.I.
Wuryandari, Desain dan Implementasi Sistem
Identifikasi Pesawat Terbang Berbasis Jaringan
Syaraf Tiruan Model Back Propagation Network,
Prosiding Seminar Nasional Teknologi Informasi
dan Aplikasinya 2009 (SENTIA09), Politeknik
Negeri Malang, Malang, 12 March 2009, pp.
F55-F60, ISSN 977-208-5234-00-7.
Nopriansyah, A.D.W. Sumari, A.I. Wuryandari,
and Andaruna, Radar Identification Friend, Foe,
or Neutral System using Aircraft’s Radar Cross
Section and Speed based on Adaptive Resonance
Theory 1 Artificial Neural Network and
Information Fusion, Proceedings of 2008
National Radar Seminar, ISSN 1979-2921, April
30, Jakarta, 81-86 (in Indonesian).
N.I. Utama, A.D.W. Sumari, and A.I.
Wuryandari, Aplikasi Jaringan Syaraf Tiruan
Adaptive Resonance Theory 1 pada Sistem
Identifikasi Pesawat Terbang, Prosiding Seminar
Nasional Teknologi Informasi dan Aplikasinya
2009 (SENTIA09), Politeknik Negeri Malang,
Malang, 12 March 2009, pp. F22-F27, ISSN
977-208-5234-00-7.
S. Haykin, Neural Networks: A Comprehensive
Foundation, IEEE Press, USA, 1994.
Download