Exeley Inc.

Guan, Yue-qi

Xu, Du

Guan, Han-qing

PLANNING OF CUTTER PATH IN HIGH SPEED MACHINING BASED ON MAPPING OF TRIANGULAR GRID SURFACE

Ju-Wei, ZHANG

Yu, WANG

Ya-le, Wang

A DEPLOYMENT ALGORITHM OF HETEROGENEOUS UNDERWATER SENSOR NETWORK BASED ON ACOUSTIC AND MAGNETIC JOINT SENSING MODEL

Jancee, B Victoria

Radha, S

Das, Nandita

ANALYSIS OF NON-BINARY FAULT TOLERANT EVENT DETECTION IN WIRELESS SENSOR NETWORKS

Xu, Liqun

Jiaming, Gong

A Fast Optical Spectrum Data Acquisition Method Based on FPGA and 
DSP 
Liqun Xu1, Gong Jiaming2 
1. 616 Atkins Rd Little Rock, AR, USA, 72211, liqunxu2000@gmail.com 
2. Xian university of Posts & Telecomm. 710061, gjm@xiyou.edu.cn 
Abstract. This paper presents a fast CCD optical spectrum data acquisition method based on FPGA, FIFO and DSP. 
Introduces a linear CCD timing sequence control signal generation and high speed ADC interface with FIFO and DSP 
in detail,  publishes this design key parts FPGA logic schematic and VHDL source code, provides a general solution for 
universal high speed CCD optical spectrum data acquisition and analysis system.  
Keywords: Linear CCD, FPGA, FIFO, DSP 
1. Introduction  
     Fiber spectrometer is the key parts in portable Raman spectrometers,  It is composed of Linear CCD detector , 
optical parts (filter, diffraction grating, lens etc) and data  acquisition,  store and processing circuit part, but how do to 
fast sample and save this optical spectrum data into PC is a critical issue for the fiber spectrometers. So this paper 
presents a powerful  method of fast CCD data acquisition, introduces the method and key parts implement hardware 
schematic and FPGA VHDL source code in detail.    
2. System design based on FPGA and DSP  
Whole system level hardware design digram is showed as figure 1, In this design, ILX511 Linear CCD is selected as 
optical spectrum detector, in order to obtain high SNR and high resulation in this system, a high performance 18 bit 
SAR ADC chip AD7641 is used,  also an asynchronized  FIFO RAM has been used as bridge between the ADC and 
DSP, FPGA is used to generate variety CCD timing clock and asynchronized  FIFO control signal. TMS320VC5509A 
DSP is used as a whole system center control unit, because of it’s fast data processing ability,low power consumption 
and built in USB 2.0 port. so by using this  DSP chip, the CCD sampling optical spectrum data can be fast transmitted 
to DSP internal SRAM and then do Raman data analyzing algorithm, also taking advantage of  this DSP USB 2.0 port, 
the sampling data can be fast sent to PC. With  DSP powerful McBSP communication port, the DSP chip can easy send 
the command to FPGA and  laser control driver. TFT color display is a option in future product update.  
A D C
Linear CCD Detector
CCD Control Signal
CCD Analogy
CCD Clock Pulse
Generator Circuit
Latch and FIFO
TMS320VC5509A DSP
Signal generator
ADC Control
Data Bus D0~D15
EXT INT0
McBSP1
Serial to Parallel
FPGA and DSP communciation
USB 2.0
Pocket PC
McBSP2
EXT INT1
Serial to Parallel
TFT Color LCD
Key Scan control IC
SDA
I2C BUS
SCL
Key
Pads
MAX7349
EP16CT144C8  FPGA
CONVST
BUSY
EMIF
Laser Control Circuit
MCU Communciation Though SPI to DSP
Output Signal
McBSP0
 
Fig 1. Raman Spectrometer System Digram 
3.  Linear CCD Time Sequence Signal Generation and Optical Spectrum Data Aacquisition  
A. ILX511 linear CCD time suquence signal generation 
The ILX511 is CCD linear image sensor, it’s feature: 2048 pixels, 14 µm x 200 µm (14 µm pitch), Single 5 V power 
supply, ultra-high sensitivity, built-in sample-and-hold circuit, maximum clock frequency is 2MHz, its wavength range 
from 400~ 1000nm, it is well fit to Raman Spectrum data analying wavelength range, when use a 785 nm laser source. 
Figure 2 shows this CCD detector time sequence diagram, figure 3 shows its analog output waveform, so in order to 
driver ILX 511 CCD , external driver circuit has to provide two time sequence control signal : one is CCD clock signal 
ΦCLK , it’s maximum frequency is 2Mhz, tpical working in 1Mhz.  another is timing driver signal ΦROG. The ΦROG 
pulse period decides the scan a frame pixel time and CCD exposure time,  since a frame has 2048 pixel + 38 dummy 
signal = 2086 clock pulse, so scaning a frame CCD pixel needs at least  2086 clocks signal period, and the true effective 
clock generated analog signal , which is sensitive optical exposure source, has only 2048 pulse. so CCD time sequence 
signal generation module has to have function as follows: 1) gernerates ΦROG pulse , ΦROG period  > 20486 X ΦCLK 
period; 2) generates more than 2086 clock pulse ΦCLK; 3) generates 2048 clock pulse with duty cycle 50% 
ΦADC_CONST, which is corrposed CCD output analog signal and synchronized to start ADC convter. The 
CCD_ROG negative pulse enable the CCD output effective and let the CCD  exposured piexl convter analog signal 
output by CCD clk clock driving . So how to create these three control pulse is the design key point.  
                       
Fig 2  ILX511 Linear CCD Clock Timing Diagram                               Fig 3 CCD Analog output 
CCD ROG
CCD CLK
ADC CONST
a frame CCD scan time sequnce
> 2088 clock pulse
2048 clock pulse
 
Fig. 4  3 key CCD control pulse waveform 
Figure 4 displays this three control pulse waveform, and figure 5 presents their FPGA loglic function schematic. In 
this design, a stable 20Mhz clock source can be obtained though FPGA internal PLL , Divide this 20 Mhz clock source 
into 1 MHz with 50% duty cycle using a divider, this 1Mhz clock function as whole CCD driving clock source, the 
CCD and ADC time sequence control signal generation module is show in Figure 5. In order to obtain a Φ ROG signal , 
use a 4 divider to get a 8 us width  pulse as a ROG generator, the ΦROG period is controlled by reciving parameter 
though McBSP port coming from DSP. 
clk_in clk4
clk_div ide_ROG
inst20
clk
rog
adc_start
ADC_Start_Ctr
inst10
AND2
inst15
NOT
inst22
ADC_CONVST
!Mhz
AND2
inst23
CCD_ROG1OUTPUT
ADC CONSTOUTPUT
CCD CLKOUTPUT
VCC
CLK In INPUT
 
Fig 5  CCD and ADC time sequence control signal generation module diagram 
In Figure 5, module inst20 implement the ΦROG pulse generation function,  module inst10 implement the CCD clk 
and ADC start pulse signal. The figure 6 shows this signal generation simulation waveforms. Inst20 and inst10 mode 
VHDL code is present in follows. the inst10 VHDL is the same with inst17 in figure 7, difference is inst17 VHDL adds 
FIFO contol function.  As figure 6 show the simulation result,  the function is exactly well meet the CCD control time 
seqence reguirment.   
 
Fig6  CCD and ADC timeing sequence generation waveform 
B. CCD data sampling and transmit  into DSP Internal SRAM  
How do the CCD sampling data fast transmit to DSP internal SRAM is secondly key design point. Since 
TMS320VC5509A  DSP has more rapidly operation speed when core clock freqency works in 200Mhz, and the ccd 
scan clcok frequnce is 1Mhz, if using the 1 Mhz pulse to directly triger the DSP external interrupt source, the high data 
rate cannot be achieved to read data from the converter. Because Interrupt latencies prevent the DSP from reacting fast 
enough and even  use a direct memory access (DMA) tramit data. since the high data tied up most of I/O port bandwith, 
the Whole DSP working efficiency will drop too much, so an ideal solution is use FIFO as buffer bridge, when whole 
2048 pixel of CCD  signal has been sampled and saved to FIFO, the bolck data can be read in burst from FIFO to DSP 
SRAM though DMA mode.   
Figure 6 give a whole CCD and FIFO control signal FPGA schematic, Lpm_counter1 is used only for  simulation 
ADC input data. Although Altera FPGA provides rich FIFO source, factual using has a lot of trick. The W_REQ and 
R_REQ has to set strictly , usually output signal W_full and R emprty signal are used as a monitor judgment, but facul  
using R_REQ as external data start to read is more better than using W_full.  from Figure 7, rdempty outpout signal 
provides a read FIFO effective signal, so use R_REQ as DSP external interrupt trigger signal, in DSP interrupt  
response routine, read whole 2048 data block, when DSP implement  read EMIF, it generates a ARE signal, which can 
be utilized as synchronized  FIFO read data clock signal R_clk;  from the Fig6 schematic, ADC_StartCtr mode provides 
ADC start signal, and ADC busy output signal is used as FIFO write data clock signal , that means ,only when ADC 
finish one conversion, the ADC conveter data is valid and the sampling data can be saved to FIFO RAM, from Fig.9 , 
Fig.10, when R_Req is ready, ADC_out data is not effective, need wait for 6 r_clk clock, so using R_Req as DSP 
interrupt source, in DSP interrupt service routine, discard first 6 data and keep reading 2048 data using DMA counter. 
So from D[0] to D[2047] total 2048 sampling data can be read to DSP SRAM.  
Fig. 8 shows the function implement simulation result of figure 6, from this waveform, the W_clk set in 10ns and 
r_clk is set 1ns, that means let r_clk is 10 times fast than W_clk , and result show read data bolck is fast 10 times than 
write data block. It sucessfully soloves the difficulity of  ADC and DSP asynchronized data read and write speed.   
VCC
ADC_IN[15..0] INPUT
VCC
R_CLK INPUT
VCC
CLK_In INPUT
CCD_CLK1OUTPUT
W_ReqOUTPUT
R_ReqOUTPUT
CCD_ROGOUTPUT
ADC_CONSTOUTPUT
Wr_f ullOUTPUT
ADC_Out[15..0]OUTPUT
W_CLKOUTPUT
rdemptyOUTPUT
ADC_In_Sim[15..0]OUTPUT
Rdusedw[11..0]OUTPUT
wrusedw[11..0]OUTPUT
clk
rog
adc_start
W_Req
R_Req
ADC_Start_Ctr
inst8
clk_in clk4
clk_div ide_ROG
inst24
AND2
inst25
NOT
inst29
AND2
inst38
AND2
inst39
AND2
inst40
up counter
clock
cnt_en
a
c
lr
q[15..0]
lpm_counter1
inst
16 bits x 4096 words
data[15..0]
w rreq
w rclk
rdreq
rdclk
aclr
w rfull
w rusedw [11..0]
q[15..0]
rdempty
rdusedw [11..0]
dcfifo2
inst26
ADC_CONVST
1 Mhz
 
Fig. 7  CCD and FIFO control signal schematic 
 
Fig.8 CCD and FIFO control signal simulatiom waveform 
 
Fig.9  Data read  initial side 
 
Fig.10.  Data read end side 
C. FPGA control mode VHDL Source Code 
------- CCD ROG signal generation VHDL  
library ieee; 
use ieee.std_logic_1164.all; 
use ieee.numeric_std.all; 
entity clk_divide_ROG is 
Port(clk_in: in std_logic; 
  clk4: out std_logic); 
end entity clk_divide_ROG; 
Architecture behave of clk_divide_ROG is 
signal clk10c: std_logic; 
Signal   div10cnt: unsigned( 15 downto 0);   
begin 
Process(clk_in) 
begin 
 if(clk_in'event and clk_in='1') then 
  div10cnt<=div10cnt+1; 
  end if; 
  if(div10cnt<="0000000000000000") then 
   clk10c<='1'; 
  elsif(div10cnt> "0000000000000001" and div10cnt< " 0001000001010100 ") then  -- more than 2088 
clock  
   clk10c<='0'; 
  elsif(div10cnt >= " 0001000001010101 ") then 
   div10cnt<="0000000000000000"; 
  end if; 
 clk4 <=   clk10c ; 
end process; 
end architecture behave; 
4. CCD, ADC , FPGA and DSP hardware Interface  
Figure 11 shows the CCD, ADC, FPGA and DSP important wires connection associated with system sampling data 
transmit. The ADC7641 ADC chip is selected, it’s an 18bit high performance SAR ADC with SNR 93 db, 2 MSPS, and 
its power supply only has 2.5 V and differential input range is ±2.5V, also it’s power consumption has only 75mw, this 
extraordinary fetures is espcially well suitable for portable and handheld Raman optical anaylis instrument. In this 
design set ADC mode in  16 bit parallel, and sampling mode set in normal  1.5 MSPS. 
TMS320VC5509A has external memory interface(EMIF) with SARM and SDRAM, and six DMA controllers, it’s 
external memory space is divided into 4 spaces by CE0~CE3, in this design, CE2 is used to select momory space from 
address 0x80,0000 to 0xc0,0000; so use 0x80,0000 to 0x0x80,0800 total 2048 byte external memory space as FIFO 
source address, and assign internal SRAM 0x 02,0000 to 0x020800 as DMA destination  memory space.  
CCD_CLK
CCD_ROG
ADC_BUSY
ADC_CONVST
ADC_Out[0..15]
R CLK
DSP_EXT_INT
ADC_IN[0..15]
9
7
6, 10
5
8
1, 2
3
4
AD7641
D[0..15]
INT0
ARE
CE2
TMS320VC5509A
analog signal OPA
FPGA
EMIF
DSP
Vout
CCD_CLKCCD_ROG
CCD Dectector
Vi+
ADC_BUSY
CNVST
ADC_D[0..15]
mode0
mode1
warp
Normal
AD7641
A D C
V C C
G N D
7
5
8
A
OR gate
  
Fig. 11    CCD, ADC , FPGA and DSP Interface                     Fig 12  Optical specturm sampling dispaly   
Figure 12 shows a 785nm laser source optical spectrum waveform, this waveform indicates that the whole system 
data sampling is success and the optical spectrum data can be fast transmitted to Pocket PC though USB 2.0.  
5. Conclusions 
Since FPGA is based on SRAM filed programmable arrary, so it can be composed of different size and depth of 
FIFO RAM. by using FPGA, variety CCD timing sequnce control signal can be flexibly creatived, for different clock 
frequency   CCD, just change the correspond conversion speed ADC chip, change FIFO reading and writing clock 
frequency, the ADC sampling data can be fast transmitted to DSP internal SRAM,  meanwhile, DSP can implement real 
time algorithm and then transmits data to PC through it’s USB port. Therfore this design provides a comprehensive 
optical spectrum data acquisition method.   
Reference 
[1] Using TI FIFO to Interface High- Speed Data Converters with TI TMS320DSPs Robert finger, TI application 
Report SDMA001 June. 2001. 
[2] Designing With TI SN74V2x5 FIFO Programmable Flags  Gary Khazan and Clayton Gibbs May 2001. 
[3] Altera Cyclone FPGA Family Datasheet, Altera corpation , May, 2008.   
[4] TMS320VC5503/5507/5509/5510 DSP Direct Memory Access (DMA) Controller Reference Guide Literature  
number SPRU587E , Jan. 2007 TI.  
[5] SONY ILX511 datasheet. 
[6] AD7641 Datasheet Analog Device Inc. 2006. 
[7] VHDL hardware description langue and digital logic circuit design, Hou Baihen, Gu Xin, 1999.  


A Fast Optical Spectrum Data Acquisition Method Based on FPGA and DSP

Lei, Zhenyu

Xin, Daxin

Hua, Jin

Motion Simulation of Bionic Hexapod Robot Based on Virtual Prototyping Technology

Zhang, Jiashan

An Efficient Density-Based Clustering Algorithm for the Capacitated Vehicle Routing Problem

Song, Yan

Yao, Xiaomei

Design of Routing Protocol and Node Structure in Wireless Sensor Network based on Improved Ant Colony Optimization Algorithm

MATSIKA, Emmanuel

PENG, Qian

CRASH SAFETY OF A TYPICAL BAY TABLE IN A RAILWAY VEHICLE

POLASIK, Jakub

WALUŚ, Konrad J.

TRANSPORT PROBLEMS                                                                                2016 Volume 11 Issue 1 
PROBLEMY TRANSPORTU                                                                 DOI: 10.20858/tp.2016.11.1.11 
 
 
barrier; impulse of force; tire, suspension; acceleration 
 
Jakub POLASIK, Konrad J. WALUŚ* 
Poznań University of Technology, Faculty of Working Machines and Transportation 
Department of Basics of Machine Design 
Piotrowo st. 3, 60-965 Poznań, Poland 
*Corresponding author. E-mail: konrad.walus@put.poznan.pl 
 
 
 
ANALYSIS OF THE FORCE DURING OVERCOMING THE ROADBLOCK – 
THE PRELIMINARY EXPERIMENTAL TESTS 
 
Summary. The paper applies to the issue of preservation of passenger vehicles during 
overcoming the roadblocks on a dry asphalt surface with the constant speed. The study 
aimed to estimate the forces which are affecting the suspension of the vehicle during 
the contact between a tyre and roadblock. Using the video recording it is possible to 
calculate the values of the tyre deflection and forces acting in the system. The obtained 
deflections of the tyres were compared to earlier laboratory test results and load-
deflection characteristics of the tyres with the same size. The roadblock with the defined 
geometry causes a dynamic deflection of the tyre. During the study a significant change 
in the geometry of the suspension has not occurred. The paper presents the characteristics 
of the vehicles research with a particular regard to model and type of tyres. 
The determination of the exact dimensions of the roadblocks and application of the ADIS 
16385 system for acquisition and archiving the features of the kinematic motion of 
the vehicle allowed to determine the speed at the moment of collision with the roadblock. 
 
 
 
ANALIZA SIŁ PODCZAS POKONYWANIA PRZESZKODY STAŁEJ – 
DOŚWIADCZALNE BADANIA PILOTAŻOWE 
 
Streszczenie. Praca dotyczy problematyki zachowania pojazdów osobowych podczas 
pokonywania przeszkody stałej na suchej nawierzchni asfaltowej, z zadaną prędkością. 
Przeprowadzone badania miały na celu oszacowanie sił oddziałujących na zawieszenie 
pojazdu w momencie styku opony z przeszkodą. Wykorzystując nagranie wideo, 
możliwe były obliczenie proporcji ugięcia opony oraz teoretyczne wyznaczenie sił 
działających w układzie. Otrzymaną wartość ugięcia opony porównano z wcześniejszymi 
wynikami badań laboratoryjnych i charakterystykami obciążenie-ugięcie dla opon o tym 
samym rozmiarze. Przeszkoda o zdefiniowanej geometrii powodowała ugięcie 
dynamiczne opony. Podczas badań nie zaobserwowano istotnej zmiany geometrii 
zawieszenia. W artykule przedstawiono charakterystyki pojazdów badawczych, ze 
szczególnym uwzględnieniem modeli i rodzaju/typu opon. Opisano dokładne wymiary 
przeszkody, a wykorzystany układ ADIS 16385 do akwizycji i archiwizacji cech 
kinematyczny ruchu pojazdu umożliwił określenie prędkości w chwili najechania na 
przeszkodę. 
 
114   J. Polasik, K.J. Waluś 
 
1. INTRODUCTION 
 
The economy is based mainly on road transport of goods and services. Intensively expanding road 
infrastructure makes it necessary to ensure the safety of all road vehicles. In many places there are 
tram tracks and crossings, which are fixed obstacles placed within the road [2, 4-6]. Additionally, 
especially in urban areas are islands bumps [3, 11], and curbs. Heavy use can lead to the formation of 
ruts faults, rifts and cracks. The resulting inequality can significantly affect the working of the tire and 
the road, which generates large peak forces acting on the tire and suspension of the vehicle. High 
acceleration values are not fully suppressed by the suspension and can be transferred to the car body 
and traveling. 
Nowadays the highways do not consists of the road surfaces only, but are also equipped with the 
infrastructure. The highway infrastructure consists of the following elements: parking bays, bus stops, 
exits, median strips, street islands, railway crossings, viaducts, bridges and all elements surrounding 
the road. The surface contaminations can be divided into two groups. The first group results from 
using of the surface, i.e. ruts, humps, irregularities of surface and destroyed surfaces. The second 
group consists of the surface road contamination due to the road traffic and dirt which was carried 
from the shoulder. During the windy weather falling branches are a high risk for the road traffic. 
Weather conditions have a high influence on the road contamination and safety of the road traffic. 
Branches, stones, unequal road subgrades or road faults can be called as the roadblocks [10]. 
The paper presents the results of the experimental tests of overcoming the roadblock by the vehicle 
with the estimated speed [7-9]. 
 
 
2. THE METHODOLOGY AND COURSE OF EXPERIMENTAL TESTS 
 
The aim of the tests was the assessment of the vehicle, tyres and suspension during overcoming 
the roadblock with the estimated velocity value of ca. 30 km/h. Two passenger cars were chosen for 
the investigations, i.e. Skoda Octavia and Volkswagen Passat (Fig. 1). Skoda Octavia estate car was 
produced in 2004. This vehicle was equipped with a front drive, six-speed manual gearbox and 
supercharged diesel engine with the capacity of 1896 cm3 and power of 96 kW. The second vehicle 
was Volkswagen Passat equipped with the same engine with power of 74 kW and five-speed manual 
gearbox. This car was produced in 2002. 
 
  
 
Fig. 1. View of the examined vehicles, Skoda (left photo), Volkswagen (right photo) 
Rys. 1. Widok badanych pojazdów, Skoda (po lewej), Volkswagen (po prawej) 
 
Both vehicles were equipped with the tyres 195/65 R15 with the steel wheel bands. Skoda was 
equipped with Gislaved Euro Frost 3 tyres manufactured with symmetrical tread in 2011. The tyres 
were very used up – this fact can be seen for the measurement results of the tread depth which are 
Analysis of the force during overcoming the roadblock… 115 
 
presented in Table 1. Volkswagen Passat was equipped with summer tyres Dunlop Sport Bluresponse 
manufactured with asymmetrical tread in 2013. The tyres had a low wear (Tab. 2) 
 
    Table 1 
Averaged measurements of the tread depth of the particular tyres for Skoda car 
 FL FR RL RR 
external 3,90 mm 5,23 mm 3,36 mm 4,01 mm 
centre 4,06 mm 4,54 mm 3,43 mm 3,90 mm 
internal 3,92 mm 4,16 mm 3,46 mm 4,01 mm 
 
Table 2 
Averaged measurements of the tread depth of the particular tyres for Volkswagen car 
 FL FR RL RR 
external 6,30 mm 5,98 mm 6,47 mm 6,30 mm 
centre 6,17 mm 6,26 mm 7,02 mm 6,59 mm 
internal 6,19 mm 6,25 mm 6,90 mm 6,74 mm 
 
The vehicles were equipped with the measuring apparatus ADIS 16385 from the Analog Devices 
company. It consisted of three acceleration sensors and three piezogyroscopes which were integrated 
in one measuring system. During the road tests the measuring apparatus was fixed on the roof of the 
vehicles. Portable computer with the dedicated software was applied for acquisition and archiving the 
measurement data. The measuring apparatus was characterised by 2% measurement uncertainty [1]. 
 
 
3. THE CHARACTERISTICS OF THE ENVIRONMENTAL CONDITIONS AND ROAD  
    SURFACE 
 
     The road tests have been done on the road with a small traffic flow along the embankment of 
the Głogowska Street near Komorniki in March 2014. This localization was chosen due to the security 
of the researchers and a lack of outsiders. The selected section of the road is characterised by a flat 
asphalt surface, lack of transverse inclination and longitudinal inclination of ca. 1%. During the tests 
the road surface was dry and clean. The road section had no defects or gaps, but inspection chambers 
were present in a regular interval. These inspection chambers did not have any influence on 
the measurement results (Fig. 2). 
 
   
 
Fig. 2. View of the investigated road section 
Rys. 2. Widok badanej drogi 
 
116   J. Polasik, K.J. Waluś 
 
During the road tests the environmental conditions were investigated before and after 
the examination of the kinematics of the vehicle motion which was overcoming the roadblock. The 
results of the environment research are presented in Tables 3 and 4. 
 
          Table 3 
The measurement results of the environmental conditions before  
the kinematic features test of the vehicle motion 
Ambient temperature 11,0 oC 
Humidity 72,0% 
Atmospheric pressure 1001,5 HPa 
Temperature of saturation point 6,1 oC 
Road surface temperature 15,4oC 
 
         Table 4 
The measurement results of the environmental conditions after  
the kinematic features test of the vehicle motion 
Ambient temperature 12,0oC 
Humidity 70,3% 
Atmospheric pressure 1001,3 HPa 
Temperature of saturation point 6,5oC 
Road surface temperature 16,2oC 
 
 
4. THE MEASURMENT RESULTS 
 
The road tests were performed from start until stop after overcoming the roadblock. The roadblock 
was placed in a distance of 30 m from the start line. The driver's task was to start, change the gear and 
overcome the roadblock with the velocity of 30 km/h. Next his task was to intensively apply brakes 
until full stop.  
A wooden board was used as the roadblock (Fig. 3). It was undercut in order to run over 
the roadblock easily. 
 
  
 
Fig. 3. View of the roadblock on the route (left photo), the dimensions of the roadblock (right picture) 
Rys. 3. Widok przeszkody stałej na drodze (po lewej), wymiary przeszkody (po prawej) 
 
Analysis of the force during overcoming the roadblock… 117 
 
Five runs over the roadblock for each vehicle were performed. The exemplary characteristics of 
the longitudinal acceleration in function of time are presented in Fig. 4. 
 
-10
-8
-6
-4
-2
0
2
4
6
8
0 2 4 6 8 10
Time [s]
A
cc
el
er
at
io
n 
[m
/s
s]
Volkswagen Passat
 
-20
-15
-10
-5
0
5
10
15
1 2 3 4 5 6 7 8 9 10
Time [s]
A
cc
el
er
at
io
n 
[m
/s
s]
Skoda Octavia
 
Fig. 4. Exemplary characteristics of the longitudinal acceleration with the visible impulse generated by the 
roadblock 
Rys. 4. Przykładowe przebiegi przyspieszeń z widocznym impulsem siły generowanym przez przeszkodę 
 
 
5. ANALYSIS OF THE MEASURMENT RESULTS 
 
The measurement tests were done for one-way traffic with uphill drive. The inclination of the road 
surface was corrected in order to obtain the real values of the longitudinal accelerations. All tests were 
additionally recorded with a camera. This allowed us to analyse the deflection of the tyre and 
suspension (frame after the frame). On the basis of the video analysis with the application of 
the software for graphics processing we have stated that the real height of the roadblock, i.e. 60 mm, 
corresponds to 22 pixels in a picture. It was calculated that 1 pixel has a height of ca. 2,72 mm. 
The pixel segmentation of the frame allowed us to determine the deflection values of tyre and 
suspension during the contact between the tyre and the roadblock (Fig. 5). This fact is presented in 
Table 5. It was assumed that the tyre deflection during the motion on a flat surface with a constant 
velocity is equal to static deflection of the tyre which is resulting from the load intensity of the wheel. 
 
118   J. Polasik, K.J. Waluś 
 
  
 
Fig. 5. View of Skoda car during overcoming the roadblock 
Rys. 5. Widok samochodu Skoda podczas pokonywania przeszkody 
 
To determine the theoretical value of the force which is generating the deflection during 
overcoming the roadblock, the laboratory tests of the relationship between the tyre deflection and 
loading for the tyre pressure of 0,2 MPa were performed (Fig. 6). After the measurements of the static 
unit pressure for the particular wheels of Skoda car, the value of the static deflection of the front right 
tyre was evaluated for loading with a static force - for the static force value of 4300 N the static 
deflection was equal to 26 mm. When we add together the value of the static deflection and the value 
of the deflection which was generated during overcoming the roadblock (36mm), then we get the total 
value of the deflection which was equal to ca. 62 mm. When we extrapolate data in accordance with 
the equation of the trend line (Fig. 6), then we get the value of the deflection force, i.e. 10266,7 N for 
the deflection value of 62 mm. 
 
         Table 5 
The deflection values of the tyre and suspension calculated on the basis of  
the quantity of pixels for Skoda car 
1 pixel = 2,72 mm 
Tyre deflection (distance between 
the wheel axis and road surface) 
Suspension deflection (distance 
between the wheel axis and fixed 
element of the body of a car) 
Before 
overcoming the 
roadblock 
During 
overcoming the 
roadblock 
Before 
overcoming the 
roadblock 
During 
overcoming 
the roadblock 
pixels mm pixels mm pixels mm pixels mm 
84 229 71 193 100 272 94 256 
Tyre deflection caused by the 
roadblock: 
13 pixels ≈ 36 mm 
Suspension deflection caused by 
the roadblock: 
6 pixels ≈ 16 mm 
 
 
Analysis of the force during overcoming the roadblock… 119 
 
 
Fig. 6. The investigations of the deflection value of the winter tyre 195/65 R15 in function of loading for the tyre 
           pressure of 0.2 MPa 
Rys. 6. Badania wartości ugięcia opony zimowej o rozmiarze 195/65 R15 w funkcji obciążenia dla ciśnienia  
              0,2 MPa 
 
 
6. CONCLUSIONS  
 
On the basis of the analysis of the forces acting on the tyre and vehicle suspension system we can 
conclude that even a small roadblock which is overcome with a low speed can generate a high impulse 
of the force. We can also notice that the force which results from overcoming the roadblock is 
completely removed by the tyre and suspension deflection. Presented investigation shows in this 
article, that force impulse depends on type of the tyres and construction of the suspension. 
The influence of suspension construction is visible in acceleration characteristics (Fig. 4). There is 
second maximum (about 23,9 s
m− ) after the first maximum of 21,8 s
m−  for Volkswagen Passat 
and 24,14 s
m−  for Skoda car. This fact has an effect on a drive comfort, the possibility of vehicle 
damage and safety of the road traffic. 
 
 
References 
 
1. Analog Devices – Technical documentation system ADIS16385. 
2. Bohatkiewicz, J. & Biernacki, S. & Drach, M. & Kozłowski, D. & Nowak, P. Zasady uspokajania 
ruchu na drogach za pomocą fizycznych środków technicznych. Opracowano na zlecenie 
Ministerstwa Infrastruktury. Umowa nr TRD/1/2008 z dnia 05.02.2008 r. 113 p. [In Polish: 
Principles of calming traffic by physical technical means, have been developed for the Ministry of 
Infrastructure]. 
3. Czajkowski, K. & Fitzgerald, S. & Foster, I. & Kesselman, C. Grid Information Services for 
Distributed Resource Sharing. In: 10th IEEE International Symposium on High Performance 
Distributed Computing. P. 181-184. IEEE Press, New York, 2001. 
4. Dębowska-Mróz, M. & Rogowski, A. & Wójcik, E. & Bator, K. Ocena wybranych elementów 
stosowanych w uspokajaniu ruchu drogowego w miastach. Autobusy. 2013. Vol. 3. P. 1665-1682. 
CD-ROM, ISSN 1509-5878. [In Polish: Evaluation of selected elements used in traffic placating 
in cities]. 
120   J. Polasik, K.J. Waluś 
 
5. Foster, I. & Kesselman, C. & Nick, J. & Tuecke, S. The Physiology of the Grid: an Open Grid 
Services Architecture for Distributed Systems Integration. Technical report. Global Grid Forum, 
2002. 
6. Foster, I. & Kesselman, C. The Grid: Blueprint for a New Computing Infrastructure. Morgan 
Kaufmann, San Francisco, 1999.  
7. Kisilowski, J. & Lozia, Z. Wybrane zagadnienia symulacji procesu hamowania samochodu 
dwuosiowego na nierównej nawierzchni drogi. AUTO – Technika motoryzacyjna. 1985. Vol. 12. 
Dodatek naukowo – techniczny [In Polish: Selected aspects of process simulation biaxial car 
braking on an uneven road surface]. 
8. Lozia, Z. Analiza ruchu samochodu dwuosiowego na tle modelowania jego dynamiki. Oficyna 
Wydawnicza PW. Warszawa, 1998. [In Polish: Traffic Analysis biaxial car on the background of 
modeling its dynamics]. 
9. Lozia, Z. & Nowakowski, J. Ocena wpływu nierówności nawierzchni drogi na skuteczność 
działania hamulców w ruchu prostoliniowym. Zeszyty Naukowe Instytutu Pojazdów. 2001. 
Vol. 1(40). P. 131-156. Warszawa. [In Polish: Assess the impact of road surface irregularities on 
the brake performance in linear motion]. 
10. Talaśka, K. & Waluś, K.J. & Berdychowski, M. Modelowanie opon samochodowych 
z wykorzystaniem oprogramowania MES (badania pilotażowe). Technika transportu szynowego. 
2012. Vol. 9. P. 3487-3494. CD-ROM [In Polish: Tire modeling using mes software (pilot study)]. 
11. Zalewski, A. Efektywność technicznych środków uspokajania ruchu w aspekcie bezpieczeństwa 
ruchu drogowego. GAMBIT 2008 VII Międzynarodowe Seminarium Bezpieczeństwa Ruchu 
Drogowego. P. 171-179. Gdańsk, 2008 [In Polish: The effectiveness of technical means of traffic 
calming in terms of road safety]. 
 
 
Received 22.11.2014; accepted in revised form 10.02.2016 


Analysis of the Force during Overcoming the Roadblock – The Preliminary Experimental Tests

Mendrek, Marlena

Grzesik, Norbert

Krzyżak, Aneta

Kuźma, Konrad

TRANSPORT PROBLEMS                                                                               2018 Volume 13 Issue 2 
PROBLEMY TRANSPORTU                                                                   DOI: 10.20858/tp.2018.13.2.3 
 
 
Keywords: fuzzy logic; fuzzy expert system; helicopter take off 
 
Marlena MENDREK, Norbert GRZESIK*, Aneta KRZYŻAK, Konrad KUŹMA 
Dęblin Air Force Academy, Department of Aviation 
303 Dywizjon Street, no. 35, 08-521, Dęblin, Poland 
*Corresponding author. e-mail: n.grzesik@wsosp.pl 
 
 
 
DIFFERENT DEFUZZIFICATION METHODS IN GUIMBAL CABRI G2 
HELICOPTER TAKEOFF POSSIBILITY EVALUATION 
 
Summary. The article discusses the aerodynamic properties of the Guimbal Cabri G2 
helicopter. The main considerations concern the helicopters ascending to the altitude 
under initial conditions, which were assumed in the project and how the change of the 
defuzzification method affects the work of the fuzzy controller. The authors analyzed the 
following: the aircraft, the crew and the payload. The authors assessed the altitude and 
the minimum time of the helicopter climb. In designing and the work simulation process, 
the authors used Matlab and Simulink software. A comparison of the influence of 
selected defuzzification methods on the work of the fuzzy controller was made. 
 
 
1. INTRODUCTION 
 
There are a lot of international publications concerning the use of fuzzy logic in the real world [5, 
9, 16], but limited in aviation (in general). Some of them describe aircraft control systems [3, 6, 8, 11, 
14, 18] and some aircraft safety issues [3, 12, 13, 15, 19]. There are also works related to the use of 
fuzzy logic to stabilize the helicopter flight [17] or to give desired horizontal velocity and to regulate 
the attitude angles, so that the helicopter achieves its desired horizontal velocities at the desired 
altitude [10]. 
Guimbal Cabri G2 (Fig. 1) is a two-seat light single-engine helicopter manufactured by the French 
company, Helicopteres Guimbal. Thanks to its auto-rotation properties, it is ranked among the safest 
helicopters worldwide. It was certified in accordance with the strictest construction standards, CS-27, 
which until now have been met only by much larger and more expensive aircrafts.  Thanks to the 
spacious cabin, low-operating costs and a large display of engine parameters of the Cabri G2, it is not 
only ideal for training inexperienced pilots, but also it proves excellent for tourist flights, 
as well as flights on demand. This helicopter is backed up by its experience in training pilots  
in ten countries around the world including Germany, Switzerland, Australia and New Zealand. 
Currently, it is being used in Dęblin.  
The rotorcraft is powered by a piston engine manufactured by Lyocoming with an electronic 
ignition system and a digital power regulator of nominal power, 180 KM reduced to 145 KM. The 
Cabri G2 is standard-equipped with modern avionics system Garmin GTM650. In place of a 
conventional tail rotor system, the designers used a Fenestron, which improves the performance, 
security, as well as enhancing the handling of the helicopter. Cabri G2 features a composite structure 
that provides extra security in the event of a hard landing [1]. 
 
28                                                                                  M. Mendrek, N. Grzesik, A. Krzyżak, K. Kuźma 
 
Table 1 
Guimbal Cabri G2 technical parameters 
  
Name Parameter 
Engine Lycoming O-360 145 KM 
Maximum Takeoff Weight 700 kg 
Empty weight 420 kg 
Max Cruising Speed 100 kt 
Cruising Speed at 85% thrust 90 kt 
Velocity Never Exceeded 130 kt 
Hover Ceiling at  Max Weight (IGE) 5000 ft 
Range (with 15 min.  final reserve fuel) 700 km 
Maximum endurance with two crew members 4,4 hr 
 
 
 
Fig. 1. Guimbal Cabri G2 (Photo by Peter von Reichenberg, www.fotostyle.cz (c) 2012) 
 
 
2. THE FUZZY EXPERT SYSTEM DESIGN FOR THE EVALUATION OF THE  
    HELICOPTER ASCENDING 
 
In the design, we used the MATLAB software. The system was designed by means of: 
1) "Fuzzy Logic Toolbox”; 
2) "Simulink”. 
The design of the system was initiated by determining the input parameters, which were called: 
• helicopter weight, 
• crew weight, 
• cargo weight. 
Each input signal is described by three Gaussian membership functions expressed in kilograms 
(Figs. 2 - 4). 
Different defuzzification methods in…                                                                  29 
 
 
 
Fig. 2. Membership function for input signal 1, "helicopter weight” [kg] 
 
 
 
Fig. 3. Membership function for input signal 2,"crew weight” [kg] 
 
 
 
Fig. 4. Membership function for input signal 3"cargo weight” [kg] 
30                                                                                  M. Mendrek, N. Grzesik, A. Krzyżak, K. Kuźma 
 
The first output signal (Fig. 5) is the altitude achievable by the aircraft. We considered the range of 
0– 1,400 ft, which corresponds to the altitudes at which the examined helicopter executes each type of 
mission (training, tourist, or special flights, most frequently en-route flights). 
 
 
 
Fig. 5. Membership function for output signal, "ascending altitude” [ft] 
 
The second output signal (Fig. 6) is “climb time” measured in minutes. This is the time in which 
the helicopter is going to rise to a given altitude. This is a parameter which additionally determines the 
safety of the executed flights because the crew is able to specify more precisely at which point of the 
flight it will be on a set-flight altitude. 
The climb time was determined in the range of 0–3.5 minutes. We used three Gaussian 
membership functions, named: slow, average, and fast. 
 
 
 
Fig. 6. Membership function for output signal, “climb time” 
 
In order to determine the correct operation of the entire system and the control plane, we 
established the deduction principles. The principles were based on the initially adopted assumptions. 
These assumptions are consistent with the principles of aerodynamics and flight dynamics of a 
helicopter flight [2]. 
Different defuzzification methods in…                                                                  31 
 
In the deduction principles, we determined the output value of the ascending altitude due to the 
helicopter weight, the weight of the crew, and the weight of the cargo, which constitute the 
construction load of the aircraft. To complete the principle base, it was necessary to determine 27 
deduction rules. Fig. 7 presents the exemplary deduction rules. 
The principle base is made up of deduction rules that specify the manner of work of the designed 
system. See examples of rules below: 
  
 
  
Fig. 7. Base of selected deduction principles 
 
 
3. ANALYSIS OF PROPER SYSTEM PERFORMANCE  
 
Having established all the required parameters, we received the control plane (Fig. 8). 
 
 
 
Fig. 8. Control surface 
 
After a thorough analysis of the control surface, it can be concluded that the results of the project 
are satisfactory with the initial assumptions.  
The table below shows the results based on 20 sample data, which were used for the analysis of the 
programme’s performance. 
In order to facilitate reading the data contained in the table, we described the data by colors. On the 
basis of the weight of the helicopter, the crew weight, and payload, as specified in the software, we 
received the resulting altitude at which the helicopter, Guimbal Cabri G2, is able to rise. The obtained 
altitude is maintained within the range of 0– 1,400 feet [ft]. The altitude range is as follows:  
32                                                                                  M. Mendrek, N. Grzesik, A. Krzyżak, K. Kuźma 
 
• Medium altitude: 500–1,000 ft; 
• High altitude: 1,000–1,500 ft. 
The samples were selected in such a way that each of the obtained altitudes corresponds to a 
similar number. 
Table 2 
Research examples for 20 samples 
 
No Helicopter weight [kg] 
Crew weight 
[kg] Payload [kg] 
 
Climb altitude 
[ft] 
 
 
Time [min]   
 
1 420 60 11 1,103 2.42 
2 100 30 5 1,115 2.85 
3 112 65 34 1,112 2.80 
4 189 22 28 1,080 2.60 
5 150 70 5 1,113 2.83 
6 170 75 32 1,080 2.69 
7 74 120 32 703 1.76 
8 200 115 18 709 1.77 
9 250 118 5 706 1.77 
10 125 116 20 707 1.77 
11 350 116 5 701 1.75 
12 530 60 18 739 1.83 
13 568 22 28 711 1.77 
14 700 20 5 700 1.75 
15 500 76 32 682 1.04 
16 300 115 33 708 1.77 
17 540 56 32 700 1.08 
18 530 118 32 347 0,87 
19 620 128 32 299 0.75 
20 568 116 5 317 0.75 
 
By analyzing the results, it is possible to observe that in the optimized system, the load weight of 
the helicopter also impacts the speed–it rises. The small load of the helicopter, the greater rising 
altitude is, with time respectively shorter than in the case of excessive load. It is essential to observe 
the weight standards due to the fact that they affect the quality and efficiency of the executed 
maneuvers, which are crucial in aviation. In addition, if the maneuvers are more efficiently made, a 
better air traffic flow is in the air. Time is a relevant factor, therefore the possibility to obtain an 
accurate estimation of the helicopter’s rising time is of valuable advantage to the system. 
 
 
4. DEFUZZIFICATION METHOD SELECTION ANALYSIS 
 
In this paper, four basic defuzzification methods were compared:  
• Center of Gravity (Centroid) 
• Center of Sums (Bisector) 
• Middle of Maximum (MOM) 
• Smallest of Maximum (SOM) 
 
On the basis of the crisp values showed in Table 2, the system calculated the crisp value on output 
signals using the four defuzzification methods listed above. The control surfaces that were obtained 
Different defuzzification methods in…                                                                  33 
 
for the results are shown in Fig. 9 to Fig. 12, and the input and output values obtained for 20 research 
examples is shown in Table 3. 
 
 
Fig. 9. Control surface – Center of Gravity  
 
 
Fig. 10. Control surface – Center of Sums 
 
 
Fig. 11. Control surface – Middle of Maximum 
34                                                                                  M. Mendrek, N. Grzesik, A. Krzyżak, K. Kuźma 
 
 
 
Fig. 12. Control surface – Smallest of Maximum 
 
Table 3 
Research examples for 20 samples with different defuzzification methods 
 
No 
Input signals Output signal - altitude 
Helicopter 
weight 
[kg] 
Crew 
weight 
[kg] 
Payload 
[kg] 
 
Centroid   
 
 
Bisector 
 
 
MOM 
 
SOM 
1 420 60 11 1,103 1,090 1,119 1,040 
2 100 30 5 1,115 1,116 1,119 1,113 
3 112 65 34 1,112 1,115 1,119 1,020 
4 189 22 28 1,080 1,110 1,119 1,080 
5 150 70 5 1,113 1,160 1,119 1,130 
6 170 75 32 1,080 1,130 1,119 1,110 
7 74 120 32 703 700 700 616 
8 200 115 18 709 700 700 574 
9 250 118 5 706 700 700 504 
10 125 116 20 707 700 700 602 
11 350 116 5 701 700 700 672 
12 530 60 18 739 714 700 588 
13 568 22 28 711 700 700 644 
14 700 20 5 700 700 700 504 
15 500 76 32 682 686 700 532 
16 300 115 33 708 700 700 588 
17 540 56 32 700 700 700 560 
18 530 118 32 347 308 210 42 
19 620 128 32 299 280 210 98 
20 568 116 5 317 280 210 140 
 
On the basis of the obtained results, the center of gravity method (centroid), center of sum method 
(bisector), and middle of maximum method (MOM) gave approximately the same results in the 
evaluation of the helicopter ascending application. For the smallest of maximum (SOM) approaches, 
there are wide variations in the results that were obtained. The reason for this is that the method uses 
the extreme smallest values for calculation of the crisp value. 
Different defuzzification methods in…                                                                  35 
 
The comparison shows that the centroid and the bisector methods are better than the MOM and 
SOM –there is more consistency in the results. 
 
 
5. THE SYSTEM WORK SIMULATION  
 
The system work simulation was conducted in the Simulink software (Fig. 13). 
The helicopter weight has a decreasing linear function assigned, as its weight during a flight always 
decreases as a result of fuel combustion. The crew weight is presented as constant and the cargo 
weight is presented in an increasing–decreasing linear function [3]. Fig. 14 presents the waveforms 
of the input signals. The results of simulation are shown in Fig. 15. 
 
 
Fig. 13. System project in Simulink software 
 
After the simulation, it was concluded that the project worked properly, because with the helicopter 
load increasing, the rising altitude diminished. In the graph, it may be deduced that the climb altitude 
depends on the load of the helicopter: the less it is loaded, the higher is the altitude; and the opposite, 
the more it is loaded, the lower the altitude it is.  
The simulation confirmed the assumption that two methods of defuzzification in the presented 
project (center of gravity and center of sum) give better results than the other two methods. 
 
 
6. SUMMARY AND CONCLUSIONS 
 
In the system designing process, the height that is possible to reach by the Guimbal Cabri G2 
helicopter in required time was determined. It was obtained with the use of the Matlab software. The 
input and output values were determined in accordance to the helicopter’s, Guimbal Cabri G2, 
technical instructions. The maximum height of 1,500 [ft] was chosen as the helicopter-cruising altitude 
or the most frequent flight altitude. However, the duration of helicopter rising was repeatedly 
measured by a pilot with 30 years of experience and equals approximately 3.5 min on an average. It 
was also possible to observe during the crew changes, that if the helicopter is under heavier load, it 
does not climb up so swiftly. This was confirmed by the project and its verification during a flight [4]. 
Designing a system with an assumed operation was strictly connected to aircraft reliability and safety 
[7].  
36                                                                                  M. Mendrek, N. Grzesik, A. Krzyżak, K. Kuźma 
 
In the simulation, the fact that the helicopter while flying consumes fuel, which effects the 
reduction of its weight, was considered. During the simulation tests, a constant weight of the crew was 
assumed. The helicopter payload changes due to fuel consumption. However, in real life, during one 
day of the flight, the crew may change and it affects the whole-helicopter weight.  
 
 
 
Fig. 14. System project input signals 
  
 
 
Fig. 15. System project output signal “ascending altitude” for different defuzzification methods 
Different defuzzification methods in…                                                                  37 
 
The four methods of defuzzification were analyzed separately according to its influence on the 
simulation results. The results obtained confirm the supposition that in the designed system the 
methods centroid and bisector will work better. 
 
 
References 
 
1. Cabri G2 Flight Manual, EASA Type certificate No R.145. Guimbal AIX-EN-PROVENCE, 
FRANCE.  
2. Bronowicz, J. Podstawy aerodynamiki śmigłowca. Świdnik, 1995. [In Polish: Basics of helicopter 
aerodynamics]. 
3. Grzesik, N. Podstawy sterowania rozmytego. Dęblin. WSOSP. 2012. ISBN 978-83-60908-73-0. 
[In Polish: Basics of fuzzy control]. 
4. Laskowski, R. Osiągi, wyważenie i planowanie lotu. Szkolenie samolotowe EASA. PILEUS. 
Styczeń, 2014. ISBN 978-83-937468-0-4. [In Polish: Performance, balance and flight planning. 
EASA aircraft training]. 
5. Anderson, G. Applying fuzzy logic in the real world. Reprinted with permission of Sensor 
magazine. Helmers Publishing. Inc. 1992. Available at: http://www.sensormag.com 
6. Doitsidis, L. & Valavanis, K.P. & Tsourveloudis, N.C. & Kontitsis, M. A framework for fuzzy 
logic based UAV navigation and control. In: Robotics and Automation. Proceedings. ICRA '04, 
Vol. 4. 2004. 
7. Jasztal, M. & Żurek, J. & Tomaszek, H. Prognozowanie uszkodzeń zagrażających bezpieczeństwu 
lotów statków powietrznych. ITE Radom. 2008. [In Polish: Forecasting damages that threaten the 
safety of aircraft flights]. 
8. Luo, J. & Lan, E. Fuzzy Logic Controllers for Aircraft Flight Control. Fuzzy Logic and Intelligent 
Systems International Series in Intelligent Technologies. 1995. Vol. 3. P. 85-124. 
9. Kacprzyk, J. & Yager, R.R. Emergency-Oriented expert systems: A fuzzy approach. Information 
Sciences. December, 1985. Vol. 37. Nos. 1–3. P. 143–155. 
10. Kadmiry, B. & Driankov, D. Fuzzy control of an autonomous helicopter. In: IFSA World 
Congress and 20th NAFIPS International Conference. 2001. 
11. Nho, K. & Agarwal, R.K. Automatic Landing System Design Using Fuzzy Logic. Journal of 
Guidance, Control, and Dynamics. 2000. Vol. 23. No. 2. P. 298-304. 
12. Lower, M. & Magott, J. & Skorupski, J. Analiza incydentów lotniczych z zastosowaniem zbiorów 
rozmytych. Prace naukowe Politechniki Warszawskiej. Transport. 2013. No. 100. P. 131-144. [In 
Polish: Analysis of aviation incidents using fuzzy sets. Scientific work of the Warsaw University of 
Technology. Transport]. 
13. Hadjimichael, M. A fuzzy expert system for aviation risk assessment. Expert Systems with 
Applications. 2009. Vol. 36. No. 3. Part 2. P. 6512-6519. 
14. Wahid, N. & Rahmat, M.F. Pitch Control System Using LQR and Fuzzy Controller. In: IEEE 
Symposium on Industrial Electronics and Applications (ISIEA 2010). Penang, Malaysia. 2010.  
P. 389-394. 
15. Demirci S. & Hajiyev, CH & Schwenke, A. Fuzzy logic based automated engine health 
monitoring for commercial aircraft. Aircraft Engineering and Aerospace Technology. 2008. 
Vol. 80. No. 5. P. 516-525. 
16. Chaudhari, S. & Patil, M. Study and Review of Fuzzy Inference Systems for Decision Making and 
Control. American International Journal of Research in Science, Technology, Engineering & 
Mathematics. December 2013-February 2014. Vol. 5(1). P. 88-92. 
17. Tanaka, K. & Ohtake, H. & Wang, H.O. A practical design approach to stabilization of a 3-DOF 
RC helicopter. IEEE transactions on control systems technology. 2004. Vol. 12(2). P. 315-325. 
 
38                                                                                  M. Mendrek, N. Grzesik, A. Krzyżak, K. Kuźma 
 
18. Usta, M.A. & Akyazi, O. & Akpinar, A.S. Aircraft roll control system using LQR and fuzzy logic 
controller. Innovations in Intelligent Systems and Applications (INISTA). 2011. DOI: 
10.1109/INISTA.2011.5946069. 
19. Żurek, J. & Grzesik, N. Fuzzy expert aircraft onboard control systems assistant. Safety reliability 
and risk analysis: Beyond the horizon. ESREL conference proceedings. London: Taylor & Francis 
Group. 2013. P. 250-251. 
 
 
Received 03.12.2016; accepted in revised form 04.06.2018 


English

DIFFERENT DEFUZZIFICATION METHODS IN GUIMBAL CABRI G2 HELICOPTER TAKEOFF POSSIBILITY EVALUATION

Not a member yet