- Categories
- Top types
- Audio & home theatre
- Cameras & camcorders
- Computer cables
- Computer components
- Computers
- Data input devices
- Data storage
- Networking
- Print & Scan
- Projectors
- Smart wearables
- Software
- Telecom & navigation
- TVs & monitors
- Warranty & support
- other →
- Top brands
- Acer
- AEG
- Aeg-Electrolux
- Canon
- Casio
- Electrolux
- Garmin
- HP
- LG
- Nikon
- Panasonic
- Philips
- Samsung
- Sony
- Yamaha
- other →
- Top types
- Infotainment
- Musical instruments
- Video games & consoles
- other →
- Top brands
- Acer
- AEG
- Asus
- Electrolux
- HP
- Juno
- LG
- Panasonic
- Philips
- Progress
- Samsung
- Sharp
- Sony
- ZANKER
- Zoppas
- other →
- Top types
- Binding machines
- Boards
- Calculators
- Correction media
- Desk accessories & supplies
- Drawing supplies
- Equipment cleansing kit
- Folders, binders & indexes
- Laminators
- Mail supplies
- Paper cutters
- Sorters
- Storage accessories for office machines
- Typewriters
- Writing instruments
- other →
- Top brands
- Bosch
- Canon
- Casio
- Craftsman
- Dell
- Epson
- Garmin
- GE
- HP
- KitchenAid
- LG
- Panasonic
- Philips
- Samsung
- Sharp
- other →
- Top types
- Bedding & linens
- Cleaning & disinfecting
- Do-It-Yourself tools
- Domestic appliances
- Home décor
- Home furniture
- Home security & automation
- Kitchen & houseware accessories
- Kitchenware
- Lighting
- other →
- Top brands
- AEG
- Aeg-Electrolux
- Bauknecht
- Bosch
- Electrolux
- HP
- Indesit
- LG
- Panasonic
- Philips
- Progress
- Samsung
- Sony
- Whirlpool
- Zanussi
- other →
- Top types
- Bags & cases
- Children carnival costumes
- Clothing care
- Clothing hangers
- Dry cleaners
- Fabric shavers
- Men's clothing
- Tie holders
- Ultrasonic cleaning equipment
- Watches
- Women's clothing
- other →
- Top brands
- Braun
- Casio
- Delta
- Garmin
- Hotpoint
- Huawei
- Indesit
- LG
- Mitsubishi Electric
- Philips
- Radio Shack
- SEVERIN
- Sony
- V7
- Whirlpool
- other →
- Top types
- Air Handlers
- Boom Lifts
- Compact Excavator
- Elevators
- Excavators
- Finishers
- Front End Loaders
- Noise Reduction Machine
- Oxygen Equipment
- Robotics
- Scrubber
- Spreader
- Tractor
- Trash Compactor
- Welding System
- other →
- Top brands
- AEG
- Aeg-Electrolux
- Bauknecht
- Canon
- Electrolux
- Garmin
- HP
- LG
- Nikon
- Panasonic
- Philips
- Samsung
- Sony
- Whirlpool
- Yamaha
- other →
- Top types
- Blood pressure units
- Electric toothbrushes
- Epilators
- Feminine hygiene products
- Foot baths
- Hair trimmers & clippers
- Makeup & manicure cases
- Men's shavers
- Personal paper products
- Personal scales
- Shaver accessories
- Skin care
- Solariums
- Teeth care
- Women's shavers
- other →
- Top brands
- AEG
- Aeg-Electrolux
- Bosch
- Canon
- Casio
- Electrolux
- Garmin
- LG
- Panasonic
- Philips
- Samsung
- Sony
- Whirlpool
- Yamaha
- Zanussi
- other →
- Top types
- Hot beverage supplies
- other →
- Top brands
- other →
- Top types
- Cars
- Electric scooters
- Motor vehicle accessories & components
- Motor vehicle electronics
- Motorcycles
- Motorhomes
- Offroad Vehicle
- Scooters
- Utility Vehicle
- other →
- Top brands
- AEG
- Aeg-Electrolux
- Bosch
- Canon
- Casio
- Electrolux
- Garmin
- GE
- LG
- Panasonic
- Philips
- Samsung
- Sony
- Yamaha
- Zanussi
- other →
- Top types
- Baby bathing & potting
- Baby furniture
- Baby safety
- Baby sleeping & bedding
- Baby travel
- Feeding, diapering & nursing
- Toys & accessories
- other →
- Top brands
- AEG
- Bosch
- Canon
- Casio
- Electrolux
- Garmin
- HP
- LG
- Panasonic
- Philips
- Samsung
- Sony
- Whirlpool
- Yamaha
- Zanussi
- other →
- Top types
- Bicycles & accessories
- Bubble machines
- Camping, tourism & outdoor
- Fitness, gymnastics & weight training
- Martial arts equipment
- Skateboarding & skating
- Smoke machines
- Sport protective gear
- Target & table games
- Water sports equipment
- Winter sports equipment
- other →
- Top brands
- Craftsman
- Daikin
- Emerson
- Epson
- Frigidaire
- Harbor Freight Tools
- HP
- Miele
- Panasonic
- Philips
- ProForm
- Samsung
- Sennheiser
- Weider
- Yamaha
- other →
- Top types
- Pet hair clippers
- other →
- Top brands
- Andis
- other →
- Top types
- Other
- Science and Education
- other →
- Top brands
- other →
- Top types
- Pill Reminder Device
- Stairlifts
- other →
- Top brands
- Alber
- Bruno
- E-PIL
- Minivator
- Savaria
- other →
![Airship Airship](/uploads/1/2/6/2/126294125/721326195.jpg)
Aug 1, 2018 - Download full-text PDF. Space Island for Conversion of Hydrogen as Energy vector). 7) Khoury, G.A., Gillet, J. Airship technology. Tesla Turbine. Tesla TurbineFull description.
2012 International Conference on Control, Automation and Information Sciences (ICCAIS)
Design and Implementation of Automatic Embedded Control Hardware and Software Systems in an Unmanned Airship Tuan Anh Nguyen, Seulki Lee and Jong-Sou Park Abstract— In order to enhance the flexibility, on-demand supply and reasonable cost of geographical probes and observations, we propose an unmanned airship system in conjunction with natural resources probe devices. The use of the unmanned airship creates favorable conditions, increases the efficiency of exploration and supports helpfully the exploitation of natural resources. This is a new approach to meet actual needs. In this paper, an automatic embedded hardware and software system of an unmanned airship are designed and implemented. This work would be a contribution to the research and development of an automatic control and embedded system on a new object. Keyword—Airship, embedded system, automatic control
I. INTRODUCTION An unmanned airship, one of the autonomous and integrated systems, is increasingly considered thanks to the rapid development of mechatronics and embedded systems. It is very useful in the “D-cube” missions, i.e. missions identified as Dangerous, Dirty or Dull [2]. With the purposes of developing a system for magnetic measurement and observation at high altitude, we started a project in design, fabrication and experiments of an unmanned airship system, one of the typical unmanned systems being considered in Research and Development (R&D). The whole system consists of a flight mechanical system, flight embedded control computer hardware and software systems, flight hardware and software systems for data acquisition, ground hardware and software systems for data collection and control. The missions of this project are to research and develop an unmanned airship system for natural resources observation from heights on any terrains, at any time. The system can significantly reduce costs, time, be utilized in highland, forest or remote areas that are difficult to probe by man power. In this paper, we introduce an automatic embedded hardware and software system as a framework for further researches on airship. In detail, dynamic and control models of the airship are constructed thoroughly. Also, the embedded hardware and software system design, control algorithm design for airship are mentioned as an extension. In the following sections, the paper contents are chronologically presented including automatic embedded control hardware and software system architecture (Section II), actual results (Section III), practical applications (Section IV), conclusions (Section V) and references.
Author profiles: Tuan Anh Nguyen, Computer Engineering Department, Korea Aerospace University, Seoul, South Korea (corresponding author; e-mail: [email protected]). Seulki Lee, Computer Engineering Department, Korea Aerospace University, Seoul, South Korea (e-mail: [email protected]). Jong-Sou Park (Prof.), Computer Engineering Department, Korea Aerospace University, Seoul, South Korea ([email protected]).
978-1-4673-0813-7/12/$31.00 ©2012 IEEE
84
II. AUTOMATIC EMBEDDED CONTROL HARDWARE AND SOFTWARE SYSTEM ARCHITECTURES
A. Airship configuration TABLE 1 AIRSHIP CONFIGURATION General mechanical configuration Length Diameter Height (H) Volume Weight Payload Gas Flight speed Maximum speed Flight altitude Engine specification
15m 3.8m 4.9m 90m3 81kg 9kg Helium 0~40km/h 75km/h 0~450m 72cc5.5hp×2 two strok 110◦ Gasoline or Oil 25:1 4hours
Maximum rotation of engine Fuel Flight time (One time fuel injection) Control surfaces FUTABA FF8 8ch Maximum operable wind speed 8m/sec Main hardware configuration FCC (Flight Control Computer) Development kit: Eddy DK v2.1 (System Base Company) Sub-CPU: ARM chips (Ublox Company) Inertial measurement unit (IMU) Crossbow NAV440 (low drift MEMS-based inertial sensors with GPS aiding) Air/Pressure sensors system Air data boom RF (Radio Frequency) Modem Products of Maxstream Company (FHSS-Frequency Hopping Spread Spectrum)
B. Practical assumptions In practice, a simple and meaningful mathematical model of the airship is required first of all, hence several assumptions are necessarily introduced. The assumptions [3] are: - The familiar aircraft dynamic modelling methods apply. - Steady slow speed rectilinear flight is assumed. - A stationary atmosphere is assumed. - Motion is described as a perturbation, not necessarily small about an initial trimmed flight condition. - The mass of the airship remains constant. - Rigid body motion only is considered, aero-elastic effects are omitted. - The airship is symmetric about the Oxz plane (defined in the following subsection), both center of buoyancy (C.B) and center of gravity (C.V) lie in that plane. The layout of the airship is classical. It has four mutually perpendicular rear fin surfaces each incorporating an aerodynamic flap type control surface and it has two independently controlled
thrust vectoring prolusion units mounted either side of the aft end of the gondola. C. Dynamics and control model of Airship As in typical aircraft practice, the flight coordinate system is defined likely same as shown in the Fig.1.
Figure 1.
Flight coordinate system illustration
A right handed orthogonal axis system is fixed in the airship and constrained to move with it. The origin O is fixed at the center of volume (C.V). The Ox axis is coincident with the symmetric axis of the airship envelope. The Oxz plane is also coincident with the longitudinal-symmetric plane of the airship. The coordinate system for attitude and position determination of the airship consists of 6 parameters: x, y, and z for position and angles Roll-Pitch-Yaw around the above axes respectively for attitude. The airship position is determined relatively with respect to the earth based on longitude and latitude. The motion of the axes is considered with respect to an initial condition which is usually trimmed equilibrium flight.
°
°
J xy = I xy + L q0005 ≡ I xy + M p0005 °
°
(5)
J yz = I yz + M r0005 ≡ I yz + N q0005 °
°
J zx = I zx + N p0005 ≡ I zx + L r0005
Unlike other solid physical objects, the balloon of airship contains gas in a giant shape compared to its overall shape; hence the center of volume (C.V) and the center of gravity (C.G) vary continuously. And as in the previous subsection, airship is assumed to be symmetric through the Oxz plane. Therefore the inertial moments along with planes Oxy and Oyz equal to zero and the acceleration along with axis Oy also equals to zero ( J xy = J yz = 0, a y = by = 0 ). We move all the forces and moments (X, Y, Z, L, M and N) impacting on the airship body to the coordinate axes. In consequence, the six degrees of freedom (6-DOF) equations of forces and moments according to the Newton’s second law of motion for each degree of freedom [6] in turn are summarized from (6) to (11) where the terms on the right hand side (RHS) are respectively components of force or moment due to aerodynamic effects, static buoyancy, gravitational force, °
aerodynamic control, and propulsion. Notice that, X u0005 =
∂X where ∂U0005
X is representative for aerodynamic force or moment. • Force Equations (respectively axial force, side force, normal force equation): mxU0005 + mz qW − m y rV + ( ma z − X q0005 ) q0005 − ...
(
)
max q 2 + r 2 − ma z pr = X a + X b + X g + X c + X p
⎛ ⎞ ⎛ ⎞ myV0005 + mx rU − ⎜ ma z + Y p0005 ⎟ p0005 + ⎜ max − Y r0005 ⎟ r0005 − mz pW +, °
(6)
°
⎝
⎠
⎝
⎠
(7)
max pq + maz qr = Ya + Yb + Yg + Yc + Yp mzW0005 + m y pV − mx qU − ⎛⎜ ma x + Z q0005 ⎞⎟ q0005 −, °
Figure 2.
2D Trimmed axes notation
Fig. 2 shows the axis notation of trimmed steady rectilinear flight where V0 denotes for total speed and Θe is body attitude. The buoyancy force B and the gravity force m.g respectively act at the center of buoyancy (C.B) and the center of gravity (C.G). They have orderly their coordinates ( bx , by , bz ) and ( ax , a y , az ) . T0 is the total engine thrust acting at a point whose precise location depends on the geometry of the installed propulsion system. So, the total velocity V0 in steady flight and the corresponding velocity components in disturbed flight are presented respectively in (1) and (2). V0 = U e + Ve + We (1) (2) Where u, v, w and p, q, r are the perturbation components of linear and angular velocities respectively with respect to trimmed equilibrium. The virtual mass and inertia effects are regarded literally as additional mass or inertial terms. The components of apparent mass, the apparent moments of inertia and the apparent products of inertia are presented from (3) to (5). mx = m − X0005 u0005 ; m y = m − Y0005v0005 ; mz = m − Z0005 w0005 ; (3) U = u + U e ;V = v + Ve ;W = w + We
°
°
°
J x = I x − L p0005 ; J y = I y − M q0005 ; J z = I z − N r0005
(4)
⎝
(
2
maz p + q
2
⎠
) + ma pr = Z x
a
(8)
+ Zb + Z g + Z c + Z p
• Moment Equations (respectively Rolling, Pitching, Yawing moment equation): J x p0005 − J xz ( r0005 + pq ) + ( J z − J y ) qr −,
⎛ ⎝
⎞ ⎠
°
− ⎜ ma z + L v0005 ⎟ V0005 − ma z ( rU − pW ) = La + Lb + Lg + Lc + Lp J y q0005 + J xz p 2 − r 2 + ( J x − J z ) pr − ⎛⎜ max + M w0005 ⎞⎟ W0005 −,
(
°
)
⎝
⎠
⎛ ⎞ max ( pV − qU ) + ⎜ maz − M w0005 ⎟ U0005 + ma z ( qW − rV ) =, °
⎝
(10)
⎠
= Ma + Mb + M g + Mc + M p J z r0005 − J xz ( p0005 − qr ) + ( J y − J x ) pq +,
⎛ ma − N° 0005 ⎞V0005 + ma ( rU − pW ) = N + N + N + N + N v ⎟ ⎜ x x a b g c p ⎝ ⎠
(11)
The terms on the right hand side of the above equations are expressed as in the equations (12) to (24) °
°
°
°
°
°
X a = X u U + X v V + X w W + X p p + X q q + X r r =, °
°
°
= X u U e + X v Ve + X w We + , °
°
°
°
°
(12)
°
+Xuu + Xv v+ Xw w+ X p p + Xq q + Xr r °
°
°
°
°
°
= Xe + X u u + X v v + X w w + X p p + X q q + X r r
85
(9)
X g + X b = − ( mg − B ) sin (θ + θ e )
(13)
Yb + Yg = ( mg − B ) sin φ cos (θ + θ e )
(14)
Z b + Z g = ( mg − B ) cos φ cos (θ + θ e )
(15)
Lb + Lg = − ( mga z + Bbz ) sin φ cos (θ + θ e )
(16)
M b + M g = − ( mgaz + Bbz ) sin (θ + θ e ) −, − ( mga x + Bbx ) cos φ cos (θ + θ e ) N b + N g = ( mgax + Bbx ) sin φ cos (θ + θ e ) °
°
(18)
°
X c = X δ ( δ e + δ r ) ; Yc = Y δ δ r ; Z c = Z δ δ e °
(17)
°
(19)
Lc = 0; M c = M δ δ e ; N c = N δ δ r X p = Ts cos μ s + Tp cos μ p = T0 cos μ
(20)
Yp = 0
(21)
Z p = −Ts sin μ s − Tp sin μ p = −T0 sin μ
(22)
L p = ( −Ts sin μ s + Tp sin μ p ) d y = δ T0 sin μ d y
(23)
M p = (Ts cos μ s + Tp cos μ p ) d z −,
− (Ts sin μ s + Tp sin μ p ) d x
(24)
perturbations φ ,θ ,ψ are also negligibly small and hence their sines and cosines take small angle approximations. By arrangements of above dynamic equations, we derive the linearized longitudinal equations of motion as (31) to (33). ⎛ ⎝
⎞ ⎠
°
mx u0005 + ⎜ ma z − X q0005 ⎟ q0005 =, °
⎛ ⎝
°
⎞ ⎠
°
°
(31)
= X e + X u u + X w w + ⎜ X q − mzWe ⎟ q + X δ ( δ e + δ r ) + , °
+ X t δ t + Te − ( mg − B )( sin θ e + θ cos θ e )
⎛ ⎞ mz w0005 + ⎜ max + Z q0005 ⎟ q0005 =, °
⎝
⎠
°
⎛ ⎝
°
⎞ ⎠
°
(32)
= Z e + Z u u + Z w w + ⎜ Z q + mxU e ⎟ q + , °
+ Z δ δ e ( mg − B )( sin θ e + θ cos θ e )
⎛ ⎝
⎞ ⎠
° ⎛ ⎞ ⎝ ⎠ ° ° ° ⎛ ⎞ = M e + M u u + M w w + ⎜ M q − ma xU e − ma zWe ⎟ q + ⎝ ⎠ °
J y q0005 + ⎜ ma z − M u0005 ⎟ u0005 − ⎜ max + M w0005 ⎟ w0005 =,
°
(33)
°
+ M t δ t + M δ δ e + Te d z − −θ {( mgaz + Bbz ) cos θ e − ( mga x + Bbx ) sin θ e } −, − ( mga z + Bbz ) sin θ e − ( mga x + Bbx ) cos θ e
In small perturbations, the trim terms sum to zero. Hence, in state space form the linearized longitudinal equations of motions are written in (34) mx0005 = ax + bu (34) x T = [ u w q θ ] u T = [δ e δ r ] (35) Figure 3.
Geometric illustration of propulsive forces and moments analysis
The unmanned airship geometric relationship of the left and right thrusts (Tp , Ts ) is similar and the tilting angles of the left and right thrusts ( μ P , μ S ) are same maneuverable direction and magnitude as shown in the Fig.3. 1 ρVT ∀C XqQ 4 1 Y = Pq∀2/3CY + ρVT ∀ ( CYp P + CYr R ) 4 1 2/ 3 Z = Pq ∀ CZ + ρVT ∀CZqQ 4 1 L = Pq ∀CL + ρVT ∀4/3 ( CYp P + CYr R ) 4 1 M = Pq∀CM + ρVT ∀4/3CMq Q 4 1 N = Pq∀CN + ρVT ∀4/3 ( CNp P + CNr R ) 4 X = Pq ∀2/ 3C X +
(25) (26) (27) (28) (29) (30)
The aerodynamic forces and moments are established with the aerodynamic wind tunnel testing as from (25) to (30), where ∀ is the volume of the airship and Pq shows the dynamic pressure. The equations of motion (6) to (12) describe fully the dynamic behavior of airship. For the sake of control, constrains of airship motion to small perturbations are made. Then, the products and squares of small perturbation variables u, v, w and p, q, r are negligible. The attitude
86
⎡ ⎛ ma − X° q0005 ⎞ mx 0 ⎜ z ⎟ ⎢ ⎝ ⎠ ⎢ ° ⎛ ⎞ ⎢ 0 mz − ⎜ max + Z q0005 ⎟ m=⎢ ⎝ ⎠ ⎢ ° ° ⎢⎛⎜ ma − M u0005 ⎞⎟ − ⎛⎜ ma + M w0005 ⎞⎟ Jy x ⎢⎝ z ⎠ ⎝ ⎠ ⎢ 0 0 0 ⎣
⎡ X° ⎢ δ ⎢ ° b = ⎢ Zδ ⎢ ° ⎢M δ ⎢ 0 ⎣ ⎡x ⎢ ⎢ ⎢z a = ⎢ ⎢ ⎢M ⎢ ⎣0 °
⎤ ⎥ ⎥ 0 ⎥ ° ⎥ Mt ⎥ ⎥ 0 ⎦ °
(37)
( (
°
xw
u
zw
°
°
(
°
u
(36)
Xt
u
°
⎤ ⎥ ⎥ 0⎥ ⎥ ⎥ 0⎥ ⎥ ⎥ 1⎦ 0
Mw
°
X q − m z We °
Z q + mxU e
) )
°
M q − ma xU e − ma z We
0
⎤ ⎥ ⎥ − ( mg − B ) sin θ ⎥ ⎥ ⎥ + Bb ) cos θ − ( mga + Bb ) sin θ } ⎥ ⎥ 0 ⎦ − ( mg − B ) cos θ e
e
)
{( mga
1
z
z
e
x
x
(38)
e
After inversing m, the equations become classical control state form as in (39) x0005 = Ax + Bu (39) x x x ⎤ x ⎤ (40) ⎡x ⎡x u
⎢z u ⎢mu ⎢ ⎣ 0
−1 A = m a = ⎢
w
q
zw
zq
mw
mq
0
1
θ
δ
⎢z ⎥ B = m − 1b = ⎢ δ ⎢mδ mθ ⎥ ⎢ ⎥ 0 ⎦ ⎣ 0 zθ ⎥
t
0 ⎥
⎥
mt ⎥
⎥
0 ⎦
Similarly, we derive the linearized lateral equations of motion in small perturbations case as in (41) to (46). mx0005 = ax + bu (41) (42) x T = [ v p r φ ] u T = [δ r ] ⎡ my ⎢ ⎢ ° ⎢ − ⎛ m a + L v0005 ⎞ ⎜ ⎟ z ⎢ m = ⎝ ⎠ ⎢ ° ⎢ ⎛⎜ m a − N v0005 ⎞⎟ x ⎢⎝ ⎠ ⎢ 0 ⎣
⎡ Y° ⎤ ⎢ δ⎥ ⎢0⎥ b=⎢ ° ⎥ ⎢N ⎥ ⎢ δ⎥ ⎣⎢ 0 ⎦⎥ ⎡Y ⎢ ⎢ ⎢L ⎢ ⎢ ⎢N ⎢ ⎢⎣ 0 °
v
°
a=
v
°
v
− ⎜ m a z + Y p0005 ⎟
⎛ ⎝
⎛ m a − Y° r0005 ⎞ ⎜ ⎟ x ⎝ ⎠
Jx
− J xz
− J xz
Jz
0
0
°
⎞ ⎠
⎤ ⎥ ⎥ 0⎥ ⎥ ⎥ 0⎥ ⎥ ⎥ 1⎦ 0
(43)
(44)
( ( (
°
Y p + m z We °
L p − ma zWe °
) ( ) ( ) (
°
Y r − m xU e °
) ) )
L r + ma zU e °
N p + ma xWe
N r − ma xU e
1
0
⎡ yv ⎢ x0005 = Ax + Bu A = m −1a = ⎢ lu ⎢ nu ⎢ ⎣0
⎤ ⎥ ⎥ − ( mga + Bb ) cos θ ⎥ ⎥ ⎥ ( mga + Bb ) cos θ ⎥ ⎥ ⎥⎦ 0
( mg
− B ) cos θ e
z
x
yp
yr
lp
lr
np
nr
0
1
z
e
x
(45)
e
yφ ⎤
⎡ yδ ⎤ ⎢ ⎥ ⎥ B=m b=⎢0⎥ ⎥ nφ ⎢ nδ ⎥ ⎥ ⎢ ⎥ 0⎦ ⎣0⎦ lφ ⎥
−1
(46)
The linearized algebraic equations (39) and (46) is used to control the airship by the control input vector u or control surfaces. The control surfaces are constituted by elevators and rudders respectively shown in Fig. 4.
The hardware architecture of autopilot airship system is briefly expressed as in the Fig. 5 (a) and Fig. 5 (b). FCC (Flight Control Computer) undertakes processing sensor data, controlling the airship. AHRS (Attitude Heading Reference System) & GPS group is in charge of detecting attitude of the airship using AGRS and acquiring position of airship using GPS signals. ADT (Air Data Terminal) is in charge of communicating with ground control system, sending flight information and receiving commands. Payload is the constituent mounted on the airship to perform the required mission. In our project, the payload is special magnetic sensor to probe the uranium resource in deep underground. Information from GPS group consists of longitude, latitude, GPS altitude, GPS heading, and GPS speed values, is periodically collected from GPS to FCS then handled and compared with on-board GPS. Output data from AHRS group consists of Roll-Pitch-Yaw angles, angular velocities p, q, r, and acceleration along with axes X, Y, Z. Those data are also collected and handled by FCS as the important inputs for control problem. Synthesizing with the mission and information sent from ground control system, FCC processor sends control signals to PWM (Pulse-wise modulation) processor. Right after that it generates pulse signals to control the actuators following the results of control problem. These actions also are described more detail in the Fig. 5(c). To increase the reliability of system, RF (Radio Frequency) and RC (Remote Control) channel communication are employed. RF channel, main and high-speed line, communicates via radio frequency. RC channel communicates via remote controller. RC is considered as the backup line in case of RF errors or failures. Two full-duplex RF modems are equipped, one on the Flight airship system and the other on the ground control system.
Figure 6.
Figure 4.
Autopilot with survey equipment
The system configuration with survey equipment shows as in the Fig. 6 in which the survey equipment is added on the main CPU. Through this architecture, we can add on or remove whatever survey equipment to/from the system without affection to whole system architecture. This is showed as the flexibility and expendability properties of the airship system architecture. 2) Flight Embedded Software The flight embedded software is arranged and organized in program, subprogram or modules. The main program consists of 7 other subprograms including SD Card, GPS, AHRS, ADC, PWM, Control, and Guidance as in the Fig. 7.
Control surfaces notation
D. Automatic embedded control system design 1) Flight Embedded Hardware
(a)
Figure 7. (c) (b) Figure 5. System Architecture of Automatic Flight Computer (a) Main devices and connections (b) Hardware Structure (c) Interface structure of data flow
87
Autopilot software structure
SD Card, GPS, AHRS are the programs communicating with peripheral devices including SD Card, GPS instruments, IMU sensors, etc. to acquire physical information of Airship system. ADC is the analog to digital converter program. PWM is the pulse-wise modulation program. And Guidance is the heading program to guide
the airship fly following the track in the mission. The Control subprogram is described in lower layer as follows. It consists of Trim Setting subprogram used to configure the Trim tabs of wings. Roll Attitude Hold, Pitch Attitude Hold and Altitude Hold are the subprograms holding the angles Roll and Pitch and the altitude of the airship. Servo Out is the subprogram controlling and monitoring the Servo motor. It receives the pulse-wise signals and deploys on the motor. And the last one is Simulator used as an automatic demonstration. 3) Automatic Control Algorithms In this paper, the complicated field on motor/engine automatic control is not going to be presented. It is supposed that in the automatic control mode, airship does not change extremely its altitude therefore the control concerned here is automatically changing the operational angles of rudder and elevator for stabilization and guidance. In practice, Yaw angle is not necessary to control because two engines fixed on gondola are able to spin around the Oz axis to provide thrust vectors. Both these engines can operate independently/dependently. In the case of favorable test environment without wind, the Yaw angle changes not so much as the airship is in automatic stabilization due to the huge body shape. Therefore we ignore controlling the Yaw angle and just consider it in guidance to change the trajectory by the measure of using another motor mounted at the airship nose to perform rotation about Oz axis. Possessing several proses such as fast response, damping improvement, maximum overshoot reduction, rise and settling time reduction, increasing bandwidth, gain and phase margin improvement, PD controller is very appropriate for airship automatic control.
principle of Pitch control. In this case, we need to control the Elevator to change the rising up or falling down of the airship head in the vertical plane. Because the Pitch change does not affect to the balance or direction, we don’t need to control other factors. Similar to Roll control, Pitch angle and the rate around axis Y are measured to be the input parameters of PD controller. In the case of altitude control as briefly illustrated in the Fig. 8 (c), the control law is almost same as Pitch control. However, we need to change the reference of control process to the Altitude Ref. GPS altitude is also collected to be one of the input parameters and compared with the altitude ref. to compute the difference to be the input of Proportional control (P). N o r th W a y P o in t
H e a d in g
E r r o r A n g le P a s s C o n d it i o n 1 P a s s C o n d it i o n 2
(a)
(b)
Figure 9. Guidance Algorithm (a) The principle of guidance algorithm (b) Derived results guided by guidance algorithm
In essence, guidance is similar to Yaw control. However, the above control algorithms of Pitch and Roll are used to stabilize the airship attitude during observation flight to obtain the stable and reliable data from specific sensors. A motor is mounted on the airship nose, it directly generates moment around the axis Z to change the Yaw angle or in other words to change the airship direction. The principle here is controlling the above motor in order to the angle error between the direction of movement measured using GPS device and the line linking current position of the airship to the way-point on the given path gradually converges. The feedback control algorithm is applied for this. This law is illustrated as in Fig 9 (a). In the Fig. 9 (b), the programmatic results of guidance following a given square-shape trajectory are shown.
(a)
III. PRACTICAL RESULTS In this section, some featured results are briefly demonstrated. The following photos in Fig. 10 show the real embedded system and the practical airship in testing. (b)
Figure 10. System configuration of automatic flight test and practical test in Zinc Mine Site
(c) Figure 8. Automatic Control Algorithm (a) Roll angle control algorithm (b) Pitch angle control algorithm (c) Altitude control algorithm
In order to control the Roll angle, we need to control the Rudder to change the airship’s inclination in its vertical plane. By that way, we can stabilize the airship around the axis X (principal axis along with the airship body). However, this change also causes the imbalance and direction change around the vertical axis; hence we employ Coordinate Turn Compensation to compensate the control signal applied on the Aileron of the control wings. By that measure, the airship can regain its balance. The angle Yaw will be compared with the Yaw reference to calculate the difference. This difference is the input for Proportional control (P). The rate around X axis is collected to be the input for Differential control (D). The above explanation is briefly illustrated as in the Fig. 8 (a). The Fig. 8 (b) illustrates the
88
For embedded software development, we used Eclipse IDE to program for automatic flight control computer. The source-codes are all in C language written for ARM processor.
Figure 11. Automatic flight control, ground control software with additional functionality
The Fig. 11 illustrates the GCS software. We added a function that displays the image of exploration area and the trajectory that the airship has been following. The middle region in this user interface, we express all real-time aerodynamic information about the flight
such as attitude, velocity, altitude, direction, target, battery energy, communication state etc. And at the right side region, we create a user manual to handle system configuration for the flight. The Fig. 12 illustrates the practical results of resource exploration real-time transmitted back from airship in flight. Those are the magnetic intensity change in the exploration area and the altitude change by time. In this figure, left side region is the control panel. It receives data arriving to COM port of control computer in GCS. It also indicates whether signals are receiving or not based on the ping mechanism. On the right side region, data of magnetic field and altitude are graphically and real-time plotted.
Figure 12. Auto zooms in/out data acquisition software with additional functionalities
Figure 13. Experiment on Zinc Mine Site
In the Fig. 13, we demonstrate a data acquisition prototype of survey equipment. The magnetic data is also illustrated in the spectral graph. It shows the magnetic density change in gradient. The recent test is shown in the Fig. 14. The exploration was performed in the highland area to find out how strong the magnetic field is there. A set of waypoints were made in order to create parallel lines on the high-view photo as a mesh. This mesh forms the trajectory with the purpose of covering as much as possible the observation area illustrated in Fig. 14 (a). A 3-D real trajectory along with the given paths is shown in Fig. 14 (b), plotted from collected data in flight.
IV. PRACTICAL APPLICATIONS The research product has been penetrating into practice with a plenty of considerable and practical applications. This highly applicable research deployed successfully in South Korea as well as in some other developing countries. A typical application of airship is such as search and rescue especially in the countries of annually natural disasters for example mine collapses, ship accidents, etc. In the case of mine collapse, airship is employed to thoroughly scan the nearby area of a collapsed mine to seek the injured workers or dead bodies underneath meters of rocks and soils. Another significant of airship application is natural resources probe or exploration. Airship possesses a very flexible and plenty of features such as low flying speed, automatic control following any predefined trajectories, flexible add-in equipment, sensors or devices, and so on. The great advantages of airship make it so considerable to be utilized in probing and exploring natural resources. Also, airship is in charge of an useful part in building up the GIS (Geographical Information System) of a country. Finally, airship has a remarkable usages in a large range of other applications such as surveillance, monitoring, exploration, warning, etc. on any terrains, at any time, as mentioned above in Dangerous, Dirty or Dull missions to replace human role. V. CONCLUSION Originally from actual demands, an automatic embedded control hardware and software system in an unmanned airship introduced in this paper are designed and fabricated to carry out several missions of mineral resources exploration. The practical results show that the system meets the given requirements of the project. This system ensures the stability and reliability to accomplish actual missions. The applications and developments of this automatic embedded control system are still fruitful for future researches and cooperation as well. This research and product would be meaningful especially for developing countries. ACKNOWLEDGEMENT This paper is a publication of the authors thanks to the support by the Basic Research Project (GP2012-006) of the Korea Institute of Geoscience and Mineral Resources (KIGAM) in cooperation with NSLab, Korea Aerospace University (KAU) and funding by the Ministry of Knowledge Economy of Korea . [1] [2]
(b) (a) Figure 14. Finally practical experiment in Zince Mine Site (a) Parallel trajectories (b) The real trajectory along with the given paths
[3] [4]
Through the Fig. 14 (b), the practical trajectory follows hard after the given one. It convinces that the process of attitude control during flight, the errors in operation and computation are acceptable. Hence, the stabilization of the airship is good enough for collecting data.
89
[5]
REFERENCES Gene F. Franklin, J. David Powell, Abbas Emami-Naeini, “ Feedback Control of Dynamic Systems”, 6th edition, Pearson Express, 2010 Enric Pastor, Juan Lopez and Pablo Royo, “A hardware/software architecture for UAV payload and mission control”, Department of Computer Architecture, Technical University of Catalonia, Castelldefels (Barcelona), Spain. Gabriel A. Khoury, J. David Gillett, “ Airship Technology”, Cambrige Aerospace Series 10, 1999 Michael V.Cook, “Flight Dynamic Principles”, 2nd edition, Elsevier Aerospace Engineering Series, 2007 Pratt, R., “Flight Control Systems: Practical Issues in Design and Implementation”, Institution of Electrical Engineers (IEEE), 2000
Design and Implementation of Automatic Embedded Control Hardware and Software Systems in an Unmanned Airship Tuan Anh Nguyen, Seulki Lee and Jong-Sou Park Abstract— In order to enhance the flexibility, on-demand supply and reasonable cost of geographical probes and observations, we propose an unmanned airship system in conjunction with natural resources probe devices. The use of the unmanned airship creates favorable conditions, increases the efficiency of exploration and supports helpfully the exploitation of natural resources. This is a new approach to meet actual needs. In this paper, an automatic embedded hardware and software system of an unmanned airship are designed and implemented. This work would be a contribution to the research and development of an automatic control and embedded system on a new object. Keyword—Airship, embedded system, automatic control
I. INTRODUCTION An unmanned airship, one of the autonomous and integrated systems, is increasingly considered thanks to the rapid development of mechatronics and embedded systems. It is very useful in the “D-cube” missions, i.e. missions identified as Dangerous, Dirty or Dull [2]. With the purposes of developing a system for magnetic measurement and observation at high altitude, we started a project in design, fabrication and experiments of an unmanned airship system, one of the typical unmanned systems being considered in Research and Development (R&D). The whole system consists of a flight mechanical system, flight embedded control computer hardware and software systems, flight hardware and software systems for data acquisition, ground hardware and software systems for data collection and control. The missions of this project are to research and develop an unmanned airship system for natural resources observation from heights on any terrains, at any time. The system can significantly reduce costs, time, be utilized in highland, forest or remote areas that are difficult to probe by man power. In this paper, we introduce an automatic embedded hardware and software system as a framework for further researches on airship. In detail, dynamic and control models of the airship are constructed thoroughly. Also, the embedded hardware and software system design, control algorithm design for airship are mentioned as an extension. In the following sections, the paper contents are chronologically presented including automatic embedded control hardware and software system architecture (Section II), actual results (Section III), practical applications (Section IV), conclusions (Section V) and references.
Author profiles: Tuan Anh Nguyen, Computer Engineering Department, Korea Aerospace University, Seoul, South Korea (corresponding author; e-mail: [email protected]). Seulki Lee, Computer Engineering Department, Korea Aerospace University, Seoul, South Korea (e-mail: [email protected]). Jong-Sou Park (Prof.), Computer Engineering Department, Korea Aerospace University, Seoul, South Korea ([email protected]).
978-1-4673-0813-7/12/$31.00 ©2012 IEEE
84
II. AUTOMATIC EMBEDDED CONTROL HARDWARE AND SOFTWARE SYSTEM ARCHITECTURES
A. Airship configuration TABLE 1 AIRSHIP CONFIGURATION General mechanical configuration Length Diameter Height (H) Volume Weight Payload Gas Flight speed Maximum speed Flight altitude Engine specification
15m 3.8m 4.9m 90m3 81kg 9kg Helium 0~40km/h 75km/h 0~450m 72cc5.5hp×2 two strok 110◦ Gasoline or Oil 25:1 4hours
Maximum rotation of engine Fuel Flight time (One time fuel injection) Control surfaces FUTABA FF8 8ch Maximum operable wind speed 8m/sec Main hardware configuration FCC (Flight Control Computer) Development kit: Eddy DK v2.1 (System Base Company) Sub-CPU: ARM chips (Ublox Company) Inertial measurement unit (IMU) Crossbow NAV440 (low drift MEMS-based inertial sensors with GPS aiding) Air/Pressure sensors system Air data boom RF (Radio Frequency) Modem Products of Maxstream Company (FHSS-Frequency Hopping Spread Spectrum)
B. Practical assumptions In practice, a simple and meaningful mathematical model of the airship is required first of all, hence several assumptions are necessarily introduced. The assumptions [3] are: - The familiar aircraft dynamic modelling methods apply. - Steady slow speed rectilinear flight is assumed. - A stationary atmosphere is assumed. - Motion is described as a perturbation, not necessarily small about an initial trimmed flight condition. - The mass of the airship remains constant. - Rigid body motion only is considered, aero-elastic effects are omitted. - The airship is symmetric about the Oxz plane (defined in the following subsection), both center of buoyancy (C.B) and center of gravity (C.V) lie in that plane. The layout of the airship is classical. It has four mutually perpendicular rear fin surfaces each incorporating an aerodynamic flap type control surface and it has two independently controlled
thrust vectoring prolusion units mounted either side of the aft end of the gondola. C. Dynamics and control model of Airship As in typical aircraft practice, the flight coordinate system is defined likely same as shown in the Fig.1.
Figure 1.
Flight coordinate system illustration
A right handed orthogonal axis system is fixed in the airship and constrained to move with it. The origin O is fixed at the center of volume (C.V). The Ox axis is coincident with the symmetric axis of the airship envelope. The Oxz plane is also coincident with the longitudinal-symmetric plane of the airship. The coordinate system for attitude and position determination of the airship consists of 6 parameters: x, y, and z for position and angles Roll-Pitch-Yaw around the above axes respectively for attitude. The airship position is determined relatively with respect to the earth based on longitude and latitude. The motion of the axes is considered with respect to an initial condition which is usually trimmed equilibrium flight.
°
°
J xy = I xy + L q0005 ≡ I xy + M p0005 °
°
(5)
J yz = I yz + M r0005 ≡ I yz + N q0005 °
°
J zx = I zx + N p0005 ≡ I zx + L r0005
Unlike other solid physical objects, the balloon of airship contains gas in a giant shape compared to its overall shape; hence the center of volume (C.V) and the center of gravity (C.G) vary continuously. And as in the previous subsection, airship is assumed to be symmetric through the Oxz plane. Therefore the inertial moments along with planes Oxy and Oyz equal to zero and the acceleration along with axis Oy also equals to zero ( J xy = J yz = 0, a y = by = 0 ). We move all the forces and moments (X, Y, Z, L, M and N) impacting on the airship body to the coordinate axes. In consequence, the six degrees of freedom (6-DOF) equations of forces and moments according to the Newton’s second law of motion for each degree of freedom [6] in turn are summarized from (6) to (11) where the terms on the right hand side (RHS) are respectively components of force or moment due to aerodynamic effects, static buoyancy, gravitational force, °
aerodynamic control, and propulsion. Notice that, X u0005 =
∂X where ∂U0005
X is representative for aerodynamic force or moment. • Force Equations (respectively axial force, side force, normal force equation): mxU0005 + mz qW − m y rV + ( ma z − X q0005 ) q0005 − ...
(
)
max q 2 + r 2 − ma z pr = X a + X b + X g + X c + X p
⎛ ⎞ ⎛ ⎞ myV0005 + mx rU − ⎜ ma z + Y p0005 ⎟ p0005 + ⎜ max − Y r0005 ⎟ r0005 − mz pW +, °
(6)
°
⎝
⎠
⎝
⎠
(7)
max pq + maz qr = Ya + Yb + Yg + Yc + Yp mzW0005 + m y pV − mx qU − ⎛⎜ ma x + Z q0005 ⎞⎟ q0005 −, °
Figure 2.
2D Trimmed axes notation
Fig. 2 shows the axis notation of trimmed steady rectilinear flight where V0 denotes for total speed and Θe is body attitude. The buoyancy force B and the gravity force m.g respectively act at the center of buoyancy (C.B) and the center of gravity (C.G). They have orderly their coordinates ( bx , by , bz ) and ( ax , a y , az ) . T0 is the total engine thrust acting at a point whose precise location depends on the geometry of the installed propulsion system. So, the total velocity V0 in steady flight and the corresponding velocity components in disturbed flight are presented respectively in (1) and (2). V0 = U e + Ve + We (1) (2) Where u, v, w and p, q, r are the perturbation components of linear and angular velocities respectively with respect to trimmed equilibrium. The virtual mass and inertia effects are regarded literally as additional mass or inertial terms. The components of apparent mass, the apparent moments of inertia and the apparent products of inertia are presented from (3) to (5). mx = m − X0005 u0005 ; m y = m − Y0005v0005 ; mz = m − Z0005 w0005 ; (3) U = u + U e ;V = v + Ve ;W = w + We
°
°
°
J x = I x − L p0005 ; J y = I y − M q0005 ; J z = I z − N r0005
(4)
⎝
(
2
maz p + q
2
⎠
) + ma pr = Z x
a
(8)
+ Zb + Z g + Z c + Z p
• Moment Equations (respectively Rolling, Pitching, Yawing moment equation): J x p0005 − J xz ( r0005 + pq ) + ( J z − J y ) qr −,
⎛ ⎝
⎞ ⎠
°
− ⎜ ma z + L v0005 ⎟ V0005 − ma z ( rU − pW ) = La + Lb + Lg + Lc + Lp J y q0005 + J xz p 2 − r 2 + ( J x − J z ) pr − ⎛⎜ max + M w0005 ⎞⎟ W0005 −,
(
°
)
⎝
⎠
⎛ ⎞ max ( pV − qU ) + ⎜ maz − M w0005 ⎟ U0005 + ma z ( qW − rV ) =, °
⎝
(10)
⎠
= Ma + Mb + M g + Mc + M p J z r0005 − J xz ( p0005 − qr ) + ( J y − J x ) pq +,
⎛ ma − N° 0005 ⎞V0005 + ma ( rU − pW ) = N + N + N + N + N v ⎟ ⎜ x x a b g c p ⎝ ⎠
(11)
The terms on the right hand side of the above equations are expressed as in the equations (12) to (24) °
°
°
°
°
°
X a = X u U + X v V + X w W + X p p + X q q + X r r =, °
°
°
= X u U e + X v Ve + X w We + , °
°
°
°
°
(12)
°
+Xuu + Xv v+ Xw w+ X p p + Xq q + Xr r °
°
°
°
°
°
= Xe + X u u + X v v + X w w + X p p + X q q + X r r
85
(9)
X g + X b = − ( mg − B ) sin (θ + θ e )
(13)
Yb + Yg = ( mg − B ) sin φ cos (θ + θ e )
(14)
Z b + Z g = ( mg − B ) cos φ cos (θ + θ e )
(15)
Lb + Lg = − ( mga z + Bbz ) sin φ cos (θ + θ e )
(16)
M b + M g = − ( mgaz + Bbz ) sin (θ + θ e ) −, − ( mga x + Bbx ) cos φ cos (θ + θ e ) N b + N g = ( mgax + Bbx ) sin φ cos (θ + θ e ) °
°
(18)
°
X c = X δ ( δ e + δ r ) ; Yc = Y δ δ r ; Z c = Z δ δ e °
(17)
°
(19)
Lc = 0; M c = M δ δ e ; N c = N δ δ r X p = Ts cos μ s + Tp cos μ p = T0 cos μ
(20)
Yp = 0
(21)
Z p = −Ts sin μ s − Tp sin μ p = −T0 sin μ
(22)
L p = ( −Ts sin μ s + Tp sin μ p ) d y = δ T0 sin μ d y
(23)
M p = (Ts cos μ s + Tp cos μ p ) d z −,
− (Ts sin μ s + Tp sin μ p ) d x
(24)
perturbations φ ,θ ,ψ are also negligibly small and hence their sines and cosines take small angle approximations. By arrangements of above dynamic equations, we derive the linearized longitudinal equations of motion as (31) to (33). ⎛ ⎝
⎞ ⎠
°
mx u0005 + ⎜ ma z − X q0005 ⎟ q0005 =, °
⎛ ⎝
°
⎞ ⎠
°
°
(31)
= X e + X u u + X w w + ⎜ X q − mzWe ⎟ q + X δ ( δ e + δ r ) + , °
+ X t δ t + Te − ( mg − B )( sin θ e + θ cos θ e )
⎛ ⎞ mz w0005 + ⎜ max + Z q0005 ⎟ q0005 =, °
⎝
⎠
°
⎛ ⎝
°
⎞ ⎠
°
(32)
= Z e + Z u u + Z w w + ⎜ Z q + mxU e ⎟ q + , °
+ Z δ δ e ( mg − B )( sin θ e + θ cos θ e )
⎛ ⎝
⎞ ⎠
° ⎛ ⎞ ⎝ ⎠ ° ° ° ⎛ ⎞ = M e + M u u + M w w + ⎜ M q − ma xU e − ma zWe ⎟ q + ⎝ ⎠ °
J y q0005 + ⎜ ma z − M u0005 ⎟ u0005 − ⎜ max + M w0005 ⎟ w0005 =,
°
(33)
°
+ M t δ t + M δ δ e + Te d z − −θ {( mgaz + Bbz ) cos θ e − ( mga x + Bbx ) sin θ e } −, − ( mga z + Bbz ) sin θ e − ( mga x + Bbx ) cos θ e
In small perturbations, the trim terms sum to zero. Hence, in state space form the linearized longitudinal equations of motions are written in (34) mx0005 = ax + bu (34) x T = [ u w q θ ] u T = [δ e δ r ] (35) Figure 3.
Geometric illustration of propulsive forces and moments analysis
The unmanned airship geometric relationship of the left and right thrusts (Tp , Ts ) is similar and the tilting angles of the left and right thrusts ( μ P , μ S ) are same maneuverable direction and magnitude as shown in the Fig.3. 1 ρVT ∀C XqQ 4 1 Y = Pq∀2/3CY + ρVT ∀ ( CYp P + CYr R ) 4 1 2/ 3 Z = Pq ∀ CZ + ρVT ∀CZqQ 4 1 L = Pq ∀CL + ρVT ∀4/3 ( CYp P + CYr R ) 4 1 M = Pq∀CM + ρVT ∀4/3CMq Q 4 1 N = Pq∀CN + ρVT ∀4/3 ( CNp P + CNr R ) 4 X = Pq ∀2/ 3C X +
(25) (26) (27) (28) (29) (30)
The aerodynamic forces and moments are established with the aerodynamic wind tunnel testing as from (25) to (30), where ∀ is the volume of the airship and Pq shows the dynamic pressure. The equations of motion (6) to (12) describe fully the dynamic behavior of airship. For the sake of control, constrains of airship motion to small perturbations are made. Then, the products and squares of small perturbation variables u, v, w and p, q, r are negligible. The attitude
86
⎡ ⎛ ma − X° q0005 ⎞ mx 0 ⎜ z ⎟ ⎢ ⎝ ⎠ ⎢ ° ⎛ ⎞ ⎢ 0 mz − ⎜ max + Z q0005 ⎟ m=⎢ ⎝ ⎠ ⎢ ° ° ⎢⎛⎜ ma − M u0005 ⎞⎟ − ⎛⎜ ma + M w0005 ⎞⎟ Jy x ⎢⎝ z ⎠ ⎝ ⎠ ⎢ 0 0 0 ⎣
⎡ X° ⎢ δ ⎢ ° b = ⎢ Zδ ⎢ ° ⎢M δ ⎢ 0 ⎣ ⎡x ⎢ ⎢ ⎢z a = ⎢ ⎢ ⎢M ⎢ ⎣0 °
⎤ ⎥ ⎥ 0 ⎥ ° ⎥ Mt ⎥ ⎥ 0 ⎦ °
(37)
( (
°
xw
u
zw
°
°
(
°
u
(36)
Xt
u
°
⎤ ⎥ ⎥ 0⎥ ⎥ ⎥ 0⎥ ⎥ ⎥ 1⎦ 0
Mw
°
X q − m z We °
Z q + mxU e
) )
°
M q − ma xU e − ma z We
0
⎤ ⎥ ⎥ − ( mg − B ) sin θ ⎥ ⎥ ⎥ + Bb ) cos θ − ( mga + Bb ) sin θ } ⎥ ⎥ 0 ⎦ − ( mg − B ) cos θ e
e
)
{( mga
1
z
z
e
x
x
(38)
e
After inversing m, the equations become classical control state form as in (39) x0005 = Ax + Bu (39) x x x ⎤ x ⎤ (40) ⎡x ⎡x u
⎢z u ⎢mu ⎢ ⎣ 0
−1 A = m a = ⎢
w
q
zw
zq
mw
mq
0
1
θ
δ
⎢z ⎥ B = m − 1b = ⎢ δ ⎢mδ mθ ⎥ ⎢ ⎥ 0 ⎦ ⎣ 0 zθ ⎥
t
0 ⎥
⎥
mt ⎥
⎥
0 ⎦
Similarly, we derive the linearized lateral equations of motion in small perturbations case as in (41) to (46). mx0005 = ax + bu (41) (42) x T = [ v p r φ ] u T = [δ r ] ⎡ my ⎢ ⎢ ° ⎢ − ⎛ m a + L v0005 ⎞ ⎜ ⎟ z ⎢ m = ⎝ ⎠ ⎢ ° ⎢ ⎛⎜ m a − N v0005 ⎞⎟ x ⎢⎝ ⎠ ⎢ 0 ⎣
⎡ Y° ⎤ ⎢ δ⎥ ⎢0⎥ b=⎢ ° ⎥ ⎢N ⎥ ⎢ δ⎥ ⎣⎢ 0 ⎦⎥ ⎡Y ⎢ ⎢ ⎢L ⎢ ⎢ ⎢N ⎢ ⎢⎣ 0 °
v
°
a=
v
°
v
− ⎜ m a z + Y p0005 ⎟
⎛ ⎝
⎛ m a − Y° r0005 ⎞ ⎜ ⎟ x ⎝ ⎠
Jx
− J xz
− J xz
Jz
0
0
°
⎞ ⎠
⎤ ⎥ ⎥ 0⎥ ⎥ ⎥ 0⎥ ⎥ ⎥ 1⎦ 0
(43)
(44)
( ( (
°
Y p + m z We °
L p − ma zWe °
) ( ) ( ) (
°
Y r − m xU e °
) ) )
L r + ma zU e °
N p + ma xWe
N r − ma xU e
1
0
⎡ yv ⎢ x0005 = Ax + Bu A = m −1a = ⎢ lu ⎢ nu ⎢ ⎣0
⎤ ⎥ ⎥ − ( mga + Bb ) cos θ ⎥ ⎥ ⎥ ( mga + Bb ) cos θ ⎥ ⎥ ⎥⎦ 0
( mg
− B ) cos θ e
z
x
yp
yr
lp
lr
np
nr
0
1
z
e
x
(45)
e
yφ ⎤
⎡ yδ ⎤ ⎢ ⎥ ⎥ B=m b=⎢0⎥ ⎥ nφ ⎢ nδ ⎥ ⎥ ⎢ ⎥ 0⎦ ⎣0⎦ lφ ⎥
−1
(46)
The linearized algebraic equations (39) and (46) is used to control the airship by the control input vector u or control surfaces. The control surfaces are constituted by elevators and rudders respectively shown in Fig. 4.
The hardware architecture of autopilot airship system is briefly expressed as in the Fig. 5 (a) and Fig. 5 (b). FCC (Flight Control Computer) undertakes processing sensor data, controlling the airship. AHRS (Attitude Heading Reference System) & GPS group is in charge of detecting attitude of the airship using AGRS and acquiring position of airship using GPS signals. ADT (Air Data Terminal) is in charge of communicating with ground control system, sending flight information and receiving commands. Payload is the constituent mounted on the airship to perform the required mission. In our project, the payload is special magnetic sensor to probe the uranium resource in deep underground. Information from GPS group consists of longitude, latitude, GPS altitude, GPS heading, and GPS speed values, is periodically collected from GPS to FCS then handled and compared with on-board GPS. Output data from AHRS group consists of Roll-Pitch-Yaw angles, angular velocities p, q, r, and acceleration along with axes X, Y, Z. Those data are also collected and handled by FCS as the important inputs for control problem. Synthesizing with the mission and information sent from ground control system, FCC processor sends control signals to PWM (Pulse-wise modulation) processor. Right after that it generates pulse signals to control the actuators following the results of control problem. These actions also are described more detail in the Fig. 5(c). To increase the reliability of system, RF (Radio Frequency) and RC (Remote Control) channel communication are employed. RF channel, main and high-speed line, communicates via radio frequency. RC channel communicates via remote controller. RC is considered as the backup line in case of RF errors or failures. Two full-duplex RF modems are equipped, one on the Flight airship system and the other on the ground control system.
Figure 6.
Figure 4.
Autopilot with survey equipment
The system configuration with survey equipment shows as in the Fig. 6 in which the survey equipment is added on the main CPU. Through this architecture, we can add on or remove whatever survey equipment to/from the system without affection to whole system architecture. This is showed as the flexibility and expendability properties of the airship system architecture. 2) Flight Embedded Software The flight embedded software is arranged and organized in program, subprogram or modules. The main program consists of 7 other subprograms including SD Card, GPS, AHRS, ADC, PWM, Control, and Guidance as in the Fig. 7.
Control surfaces notation
D. Automatic embedded control system design 1) Flight Embedded Hardware
(a)
Figure 7. (c) (b) Figure 5. System Architecture of Automatic Flight Computer (a) Main devices and connections (b) Hardware Structure (c) Interface structure of data flow
87
Autopilot software structure
SD Card, GPS, AHRS are the programs communicating with peripheral devices including SD Card, GPS instruments, IMU sensors, etc. to acquire physical information of Airship system. ADC is the analog to digital converter program. PWM is the pulse-wise modulation program. And Guidance is the heading program to guide
the airship fly following the track in the mission. The Control subprogram is described in lower layer as follows. It consists of Trim Setting subprogram used to configure the Trim tabs of wings. Roll Attitude Hold, Pitch Attitude Hold and Altitude Hold are the subprograms holding the angles Roll and Pitch and the altitude of the airship. Servo Out is the subprogram controlling and monitoring the Servo motor. It receives the pulse-wise signals and deploys on the motor. And the last one is Simulator used as an automatic demonstration. 3) Automatic Control Algorithms In this paper, the complicated field on motor/engine automatic control is not going to be presented. It is supposed that in the automatic control mode, airship does not change extremely its altitude therefore the control concerned here is automatically changing the operational angles of rudder and elevator for stabilization and guidance. In practice, Yaw angle is not necessary to control because two engines fixed on gondola are able to spin around the Oz axis to provide thrust vectors. Both these engines can operate independently/dependently. In the case of favorable test environment without wind, the Yaw angle changes not so much as the airship is in automatic stabilization due to the huge body shape. Therefore we ignore controlling the Yaw angle and just consider it in guidance to change the trajectory by the measure of using another motor mounted at the airship nose to perform rotation about Oz axis. Possessing several proses such as fast response, damping improvement, maximum overshoot reduction, rise and settling time reduction, increasing bandwidth, gain and phase margin improvement, PD controller is very appropriate for airship automatic control.
principle of Pitch control. In this case, we need to control the Elevator to change the rising up or falling down of the airship head in the vertical plane. Because the Pitch change does not affect to the balance or direction, we don’t need to control other factors. Similar to Roll control, Pitch angle and the rate around axis Y are measured to be the input parameters of PD controller. In the case of altitude control as briefly illustrated in the Fig. 8 (c), the control law is almost same as Pitch control. However, we need to change the reference of control process to the Altitude Ref. GPS altitude is also collected to be one of the input parameters and compared with the altitude ref. to compute the difference to be the input of Proportional control (P). N o r th W a y P o in t
H e a d in g
E r r o r A n g le P a s s C o n d it i o n 1 P a s s C o n d it i o n 2
(a)
(b)
Figure 9. Guidance Algorithm (a) The principle of guidance algorithm (b) Derived results guided by guidance algorithm
In essence, guidance is similar to Yaw control. However, the above control algorithms of Pitch and Roll are used to stabilize the airship attitude during observation flight to obtain the stable and reliable data from specific sensors. A motor is mounted on the airship nose, it directly generates moment around the axis Z to change the Yaw angle or in other words to change the airship direction. The principle here is controlling the above motor in order to the angle error between the direction of movement measured using GPS device and the line linking current position of the airship to the way-point on the given path gradually converges. The feedback control algorithm is applied for this. This law is illustrated as in Fig 9 (a). In the Fig. 9 (b), the programmatic results of guidance following a given square-shape trajectory are shown.
(a)
III. PRACTICAL RESULTS In this section, some featured results are briefly demonstrated. The following photos in Fig. 10 show the real embedded system and the practical airship in testing. (b)
Figure 10. System configuration of automatic flight test and practical test in Zinc Mine Site
(c) Figure 8. Automatic Control Algorithm (a) Roll angle control algorithm (b) Pitch angle control algorithm (c) Altitude control algorithm
In order to control the Roll angle, we need to control the Rudder to change the airship’s inclination in its vertical plane. By that way, we can stabilize the airship around the axis X (principal axis along with the airship body). However, this change also causes the imbalance and direction change around the vertical axis; hence we employ Coordinate Turn Compensation to compensate the control signal applied on the Aileron of the control wings. By that measure, the airship can regain its balance. The angle Yaw will be compared with the Yaw reference to calculate the difference. This difference is the input for Proportional control (P). The rate around X axis is collected to be the input for Differential control (D). The above explanation is briefly illustrated as in the Fig. 8 (a). The Fig. 8 (b) illustrates the
88
For embedded software development, we used Eclipse IDE to program for automatic flight control computer. The source-codes are all in C language written for ARM processor.
Figure 11. Automatic flight control, ground control software with additional functionality
The Fig. 11 illustrates the GCS software. We added a function that displays the image of exploration area and the trajectory that the airship has been following. The middle region in this user interface, we express all real-time aerodynamic information about the flight
such as attitude, velocity, altitude, direction, target, battery energy, communication state etc. And at the right side region, we create a user manual to handle system configuration for the flight. The Fig. 12 illustrates the practical results of resource exploration real-time transmitted back from airship in flight. Those are the magnetic intensity change in the exploration area and the altitude change by time. In this figure, left side region is the control panel. It receives data arriving to COM port of control computer in GCS. It also indicates whether signals are receiving or not based on the ping mechanism. On the right side region, data of magnetic field and altitude are graphically and real-time plotted.
Figure 12. Auto zooms in/out data acquisition software with additional functionalities
Figure 13. Experiment on Zinc Mine Site
In the Fig. 13, we demonstrate a data acquisition prototype of survey equipment. The magnetic data is also illustrated in the spectral graph. It shows the magnetic density change in gradient. The recent test is shown in the Fig. 14. The exploration was performed in the highland area to find out how strong the magnetic field is there. A set of waypoints were made in order to create parallel lines on the high-view photo as a mesh. This mesh forms the trajectory with the purpose of covering as much as possible the observation area illustrated in Fig. 14 (a). A 3-D real trajectory along with the given paths is shown in Fig. 14 (b), plotted from collected data in flight.
IV. PRACTICAL APPLICATIONS The research product has been penetrating into practice with a plenty of considerable and practical applications. This highly applicable research deployed successfully in South Korea as well as in some other developing countries. A typical application of airship is such as search and rescue especially in the countries of annually natural disasters for example mine collapses, ship accidents, etc. In the case of mine collapse, airship is employed to thoroughly scan the nearby area of a collapsed mine to seek the injured workers or dead bodies underneath meters of rocks and soils. Another significant of airship application is natural resources probe or exploration. Airship possesses a very flexible and plenty of features such as low flying speed, automatic control following any predefined trajectories, flexible add-in equipment, sensors or devices, and so on. The great advantages of airship make it so considerable to be utilized in probing and exploring natural resources. Also, airship is in charge of an useful part in building up the GIS (Geographical Information System) of a country. Finally, airship has a remarkable usages in a large range of other applications such as surveillance, monitoring, exploration, warning, etc. on any terrains, at any time, as mentioned above in Dangerous, Dirty or Dull missions to replace human role. V. CONCLUSION Originally from actual demands, an automatic embedded control hardware and software system in an unmanned airship introduced in this paper are designed and fabricated to carry out several missions of mineral resources exploration. The practical results show that the system meets the given requirements of the project. This system ensures the stability and reliability to accomplish actual missions. The applications and developments of this automatic embedded control system are still fruitful for future researches and cooperation as well. This research and product would be meaningful especially for developing countries. ACKNOWLEDGEMENT This paper is a publication of the authors thanks to the support by the Basic Research Project (GP2012-006) of the Korea Institute of Geoscience and Mineral Resources (KIGAM) in cooperation with NSLab, Korea Aerospace University (KAU) and funding by the Ministry of Knowledge Economy of Korea . [1] [2]
(b) (a) Figure 14. Finally practical experiment in Zince Mine Site (a) Parallel trajectories (b) The real trajectory along with the given paths
[3] [4]
Through the Fig. 14 (b), the practical trajectory follows hard after the given one. It convinces that the process of attitude control during flight, the errors in operation and computation are acceptable. Hence, the stabilization of the airship is good enough for collecting data.
89
[5]
REFERENCES Gene F. Franklin, J. David Powell, Abbas Emami-Naeini, “ Feedback Control of Dynamic Systems”, 6th edition, Pearson Express, 2010 Enric Pastor, Juan Lopez and Pablo Royo, “A hardware/software architecture for UAV payload and mission control”, Department of Computer Architecture, Technical University of Catalonia, Castelldefels (Barcelona), Spain. Gabriel A. Khoury, J. David Gillett, “ Airship Technology”, Cambrige Aerospace Series 10, 1999 Michael V.Cook, “Flight Dynamic Principles”, 2nd edition, Elsevier Aerospace Engineering Series, 2007 Pratt, R., “Flight Control Systems: Practical Issues in Design and Implementation”, Institution of Electrical Engineers (IEEE), 2000