First International Conference on Emerging Trends in Engineering and Technology

Longitudinal Robust Stability Augmentation for Micro air Vehicle - Design

and Validation

Dr. M. Meenakshi1 and Prof. .M Seetharama Bhat2

1. Professor, Dept. of Instrumentation Technology, Dr. AIT Bangalore, India
Email : meenakshi_mbhat@yahoo.com
2. Professor, Dept. Of Aerospace Engineering, I.I.Sc. Banaglore

Abstract remote or dangerous to send human scouts. Operational

MAVs will enable a number of important missions
This paper presents a robust fixed order H2 which include chemical/radiation spill monitoring,
controller designed using Strengthened discrete forest-fire reconnaissance, visual monitoring of
optimal projection equations (SDOPEs), which volcanic activity, surveys of natural disaster areas, and
approximate the first order necessary optimality even inexpensive traffic and accident monitoring [1, 2]
condition. The novelty of this work is the application of etc. Additional on-board sensors can further augment
the robust H2 controller to a micro aerial vehicle MAV, MAV mission profiles to include, for example airborne
named Sarika2 developed in house. The controller is chemical analysis.
designed in discrete domain for the longitudinal In military, one of the primary roles for MAVs will
dynamics of Sarika2 in the presence of low frequency be as small-unit battlefield surveillance agents, where
atmospheric turbulence (gust) and high frequency they act as an extended set of eyes in the sky. This use
sensor noise. The design specification includes of MAV technology is intended to reduce the risk to
simultaneous stabilization, disturbance rejection and military personnel. MAVs can penetrate potential
noise attenuation over the entire flight envelope of the opposite camps and other targets prior to any action
vehicle. Six degrees of freedom nonlinear equations against on those targets, which are virtually
are linearized, and validated using nonlinear undetectable from the ground. This significantly raises
simulation model and is used for the controller design. the chance for overall mission success.
The resulting controller performance is Researchers at the Aerospace Engineering
comprehensively analyzed by means of linear and Department at the Indian Institute of Science (I.I.Sc)
nonlinear simulation. Bangalore, India have established a long track record
in designing, building and test-flying (remotely human-
1. Introduction piloted) practical MAVs [3-4]. Fig. 1 shows one among
few of the recently developed MAVs as well as small
Ever since mankind’s first powered flight was made, UAVs at I.I.Sc. While much progress has been made in
research efforts have continuously pushed the envelope the design of ever-smaller MAVs by researchers at
to create flying machines that are faster and/or larger I.I.Sc, India and others in the past five years, there is a
than ever before. Now however; there is an effort to significant scope for research on autonomous MAV so
design aircraft at the other end of largely unexplored as to improve their utility to wide array of missions.
spectrum, where the desire for low cost portable, low- The first challenging step in achieving such MAV
altitude aerial surveillance has driven the development autonomy is the design and development of a robust
and testing of aircraft that are as small as possible — in flight stabilization system because, the uncertainties in
other words, on the scale and in the operational range the mathematical model associated with the low
of small birds. Vehicles in this class of small-scale Reynolds number flight are not fully understood and is
aircraft are known as Micro Air Vehicles or MAVs high. MAVs have very low moments of inertial
Equipped with small video cameras and property; hence they are highly vulnerable to rapid
transmitters, MAVs have great potential for angular accelerations.. Another potential source of
surveillance and monitoring tasks in areas either too instability for MAV is the relative magnitudes of wind
gusts, which are much higher at the MAV scale than for

978-0-7695-3267-7/08 $25.00 © 2008 IEEE 771

DOI 10.1109/ICETET.2008.213

Authorized licensed use limited to: GE GLOBAL RESEARCH. Downloaded on June 19, 2009 at 08:26 from IEEE Xplore. Restrictions apply.
 larger aircraft [1]. An average wind gust can 2. Longitudinal Dynamics of Sarika 2
immediately effect a dramatic change in the flight path Sarika2 (Fig.1) is a remotely piloted small flying
of these vehicles. In addition, the pilot may find it vehicle with a wing span of about 0.6m and 0.8m
difficult to control the aircraft based on visual cues, if length. The vehicle is weighting about 1.75k.g at its
its dynamic modes are of high frequency and are takeoff. The control surfaces are outboard elevators,
lightly damped. Hence, robust flight controller plays an inboard ailerons and rudder. The power plant is a 4cc
important role to simplify the task of operating the propeller engine with methanol plus castor oil as fuel.
MAV while enhancing the utility of MAVs for a wide Sarika2 can carry a payload comprising video camera,
range of missions. Reduced fixed order, robust H2 sensors, and transmission systems. Sarika2 has a
control is an attractive option among the several robust swept-back delta wing without a horizontal tail. It has a
multivariable methods for controller design, as H2 planform area of 0.195sqm, and a constant area square
norm is the more realistic measure of the performance section fuselage of width 0.06m.
[5]. This allows the direct incorporation of the The nonlinear six degrees of freedom equations of
multivariable robustness measures in to the the vehicle is given by the following force balance and
optimization criterion. Reduced fixed order controller moment balance equations:
can be synthesized in one step in contrast to the other Force Equations
multivariable robust controller design methods F
Uɺ 1 = RV1 − QW1 − g 0 sin θ + x
involving two step designs, and hence, guarantees the m
robustness and stability [6]. Therefore, to deal with the (1a)
Fy
above mentioned challenges in MAV flight control, Vɺ1 = − RU 1 + PW1 + g 0 sin φ cos θ +
m
this paper proposes this paper aims at the design of F
digital longitudinal stability augmentation system to Wɺ 1 = QU 1 − PV1 + g 0 cos φ cos θ + z
m
improve the handling qualities of Sarika-2 and to Moment Equations
achieve satisfactory closed loop disturbance rejection, Pɺ = (c1 R + c 2 P )Q + c3 L + c 4 N
sensor noise attenuation and robustness specifications.
Qɺ = c PR − c ( P 2 − R 2 ) + c M (1b)
Since Sarika-2 use non inertial quality sensors, and the 5 6 7
velocity sensor is not available, a single controller is Rɺ = (c8 P − c 2 R )Q + c 4 L + c9 N
where, c = (J y − J z )J z − J xz ; c = (J x − J y + J z )J xz ;
designed at the central operating point of the vehicle to 2

achieve the desired specifications over the entire flight 1 2

envelope. Controller is designed using literalized J x J z − J xz
2
J x J z − J xz
2

model and is validated using nonlinear simulation Jz ;c = J xz ,c = Jz − Jx ,

c3 = 4
model of Sarika2. J x J z − J xz2 J x J z − J xz2 5 Jy
J xz
c6 = ;c = 1 ;
Jy 7 Jy
J x (J x − J y ) + J xz2 ;
Jx
c8 = c9 =
J x J z − J xz2
J x J z − J xz2
The following two equations (eq. 1c and 1d) gives
the kinematic and navigations equations of the aircraft,
which is needed to develop the complete nonlinear
simulation model of the aircraft.
Figure 1. Micro Air Vehicle: Sarika-2 Kinematic equations
φɺ = P + tan θ (Q sin φ + R cos φ )
(1c)
The organization of this paper is as follows. Section θɺ = Q cos φ − R sin φ
2 briefly explains longitudinal dynamics of Sarika 2
Q sin φ + R cos φ
and its verification using nonlinear simulation model. ψɺ =
cosθ
Next section 3 presents briefly about the controller
design methodology. Section 4 describes the simulation
results using linear and nonlinear simulation models.
Finally conclusions are drawn in section 5

772

Authorized licensed use limited to: GE GLOBAL RESEARCH. Downloaded on June 19, 2009 at 08:26 from IEEE Xplore. Restrictions apply.
 Navigation Equations Table 2 gives the steady state coefficients for each
pɺ n = U 1 cosθ cosψ + V1 (− cos φ sin ψ + sin φ sin θ cosψ ) trimmed flight condition calculated using the analytical
+ W1 (sin φ sin ψ + cos φ sin θ cosψ )
(1d)
relations. For comparison, Table 2 also lists the
pɺ e = U 1 cos θ sin ψ + V1 (cos φ cosψ + sin φ sin θ sinψ ) corresponding values obtained from nonlinear
+ W1 (− sin φ cosψ + cos φ sin θ sin ψ ) simulation [9].
hɺ = U 1 sin θ − V1 sin φ cosθ − W1 cos φ cosθ Table2. Trim values and Steady state Coefficients
For the controller design purpose, the above set of Speed CL1 CD1
nonlinear equations is decoupled using small m/s.
perturbation theory and developed linearized Analytical Nonlinear Analytical Nonlinear
longitudinal and lateral dynamics of the vehicle. Approach Model Approach Model
Linearized longitudinal state space equations (2) are 16 0.56 0.5 0.14 0.104
developed for a straight and level flight, trimmed at six 18 0.43 0.4 0.10 0.075
operating points in the speed range of 16 - 26 m/s. The 20 0.36 0.33 0.08 0.06
state variables are
x = [ ∆u α q θ ]T
22 0.29 0.27 0.07 0.058
24 0.23 0.21 0.06 0.055
∆uɺ = ( X u + X Tu )∆u + X α (α +
wg Z
) + u q − g (cos θ1 )θ + X δeδ e
(2a)
U1 U1
0.20 0.19 0.05 0.053
Zu Z wg  Zq  Z (2b)
αɺ = ∆ u + α (α + ) +  + 1  q + δe δ e The continuous state space model is discretized at
U1 U1 U1  U1  U1
50 Hz (to synchronize with the command PWM input
wg (2c)
qɺ = (M u + M Tu ) ∆u + (M α + M Tα )(α + ) + M q q + M δ eδ e received at the vehicle from radio/pilot command from
U1
ground station).The final linearized model used for the
θɺ = q (2d) controller synthesis includes one sampling period delay
where, wg = α is the angle of attack due to vertical to account for computational time requirements.
g
U1 Hence, the final model of the plant consists of six states
wind gust wg. The elevator is actuated by electro- (four for MAV airframe, one for actuator and one for
mechanical servo systems. The dynamics of the servo delay), one control input, wind disturbance input and
actuator measured experimentally, is given by, two sensor outputs from rate gyro and accelerometers.
δɺe = − 9.5δ e + 6.37 u (3)
3. Fixed order H2 Controller Design
The measured variables are normal acceleration and Fixed reduced order H2 controller is designed to
pitch rate of the vehicle. The normal acceleration at the meet the following closed loop specifications. The
centre of gravity c.g. of the vehicle is given by, main requirement of stability augmentation system
a z = U 1 (αɺ − q ) (4) (SAS) is towards improvement of handling qualities
The coefficients of the equations (2a – 2d), known summarized as in S1.
as aerodynamic stability derivatives are computed S1: Level - 1-flying qualities of stable airframe
using analytical approach [7,8] and are refined using dynamics:
wind tunnel generated data [3]. The computed values Short period damping ratio: 0.35≤ ξsp≤ 1.3
are again validated using nonlinear simulation model, Phugoid damping ratio: ξp ≥ 0. 5
named FAST [9]. Table 1 shows that the trim values S2: Disturbance rejection Specification: Minimize
calculated by analytical means match well with those the sensitivity function below 0 dB for ω < 9 rad/sec.
obtained by means of nonlinear simulation. S3: Sensor noise attenuation Specification: Obtain
Table 1. Trim settings of alpha and elevator –40dB/decade roll off above ω = 20rad/sec.
Trim Elevator δe in Trim Alpha α in S4: Robustness Specification:
Speed Degrees Degrees The controller should be robust to structured and
m/Sec. Analytical Nonlinear Analytical Nonlinear unstructured uncertainty in plant models at all flight
Approach Model Approach Model conditions Apart from the above specifications, the
closed loop system should also be robust to maximum
16 -16.02 -15.91 16.93 17.96 expected time delays, which may arise due to
18 -15.21 -15.17 13.24 13.26 computational complexity. In addition, the control
20 -14.68 -14.42 10.59 10.54 surface deflection should not exceed its full-scale
22 -13.55 -13.36 8.64 8.67 deflection of +16degrees.
24 -12.69 -12.57 7.15 7.29 To meet the above closed loop requirements, fixed
26 -12.02 -11.96 5.99 6.23 order H2 controller is designed [10] by considering the

773

Authorized licensed use limited to: GE GLOBAL RESEARCH. Downloaded on June 19, 2009 at 08:26 from IEEE Xplore. Restrictions apply.
 performance objective of minimization of H2 norm of initial conditions for controller can safely be set to
the closed loop transfer function Tzw given by: zero.
W1 ( z )S 0 ( z )  (5)
In order to asses the performance of the controller,
min || Tzw ||= min  
W2 (z )S i ( z )K ( z ) 2
two different conditions are considered. First the
aircraft is trimmed for different cruise speeds at
where, w is the Tzw is the transfer function between constant altitude and the closed loop response to
the performance outputs i.e output sensitivity S0 doublet input is simulated. Next, the aircraft is trimmed
( S o = (I + GKK1 )−1 )and control sensitivity SiK (with at different altitudes and closed loop response to
S i = (I + KK1G ) ) functions to the disturbance input. doublet input is simulated.
−1

4.1.1 Closed loop Response due to variation in

Also K represents the controller transfer function. W1
Cruise Speed
and W2 are the weighting matrices used to minimize The closed loop nonlinear responses are
sensitivity and control sensitivity at low frequency and simulated at 3 different nominal flight speeds, (16
high frequency specified in the design specifications. m/sec, 20 m/s and 26 m/s) at constant altitude. After
Reduced fixed order controller is designed by solving the aircraft is trimmed for straight and level flight, the
strengthened discrete optimal projection equations. simulation is started at t = 0 with a doublet input along
4. Results and Analysis the pitch axis. i.e., a positive 0.1 ms input is applied at
Suitable weighing matrices are selected by trial and t = 0 and held it for 2 s. Next, a -0.1 ms. input is
error method and minimum order controller is designed applied and is zeroed after 2 s. The closed loop
to meet the design specifications. It is found that a third responses obtained from nonlinear simulation models,
order controller is sufficient to meet the closed loop at different flight conditions, are given in Figs. 2 – 4
requirements. The two elements of transfer function .
matrix of the third order controller K1×2 is given by,
K (z ) =
1
∆
[
0.0019(z − 0.1614 )(z − 0.9904 ) (
0.036 z 2 − 2.005 z + 1.008 )]
where, ∆ = ( z − 0.3828)( z − 0.9397 )(z − 0.9875) .
With this controller, at all flight conditions, short
period and phugoid damping remains greater than 0.35
and 0.5 respectively, hence, the stringent level-one
flying quality requirement is met.
4.1 Nonlinear simulation results
In practice, it is very difficult to fly the model
continuously at constant altitude and to initiate control
action from an equilibrium condition as assumed while
designing the controller. The variation in altitude in
real flight can give rise to variations in density, which
in turn can affect the dynamics of the aircraft. Also, the
effect of the coupling of longitudinal and lateral
dynamics can play a major role on the aircraft
performance. In order to assess the performance of the
controller in the presence of such uncertainties, full
scale closed loop nonlinear simulation with both
longitudinal controller is presented in this section. The
nonlinear program used for this purpose is the FAST
package developed by NAL, Bangalore [9]. FAST is
only an open loop simulation software package for the
given aircraft. In order to simulate the closed loop
response of Sarika-2, at first the input file is formulated
in the standard format [9]. In order to simplify the
controller implementation procedure in real time, Figure. 2a Closed Loop Responses from Nonlinear and Linear
controller is placed in the feedback path, in contrast to Models at 20m/s
the cascade configuration used while designing the
controller. This is because, in feedback configuration,

774

Authorized licensed use limited to: GE GLOBAL RESEARCH. Downloaded on June 19, 2009 at 08:26 from IEEE Xplore. Restrictions apply.
 Figure. 2b Closed Loop Responses from Nonlinear and Linear
Models at 20m/s

Figure. 4. Closed Loop Responses from Nonlinear and Linear

Models at 26m/s

Closed loop responses simulated using linear

models are also included in the same figures for the
sake of comparison. At 18 m/s, responses from the
nonlinear simulation are more oscillatory compared to
linear simulation responses. However, at 20 m/s and 26
m/s, the responses from nonlinear and linear
simulations match well. Figs. 2c- 4c show the trajectory
of the height variation at different flight conditions.
Responses indicates that, that straight and level flight is
maintained at all flight speeds, since, the gain/loss of
Figure. 3 Closed Loop Responses from Nonlinear and Linear height is very small (at 20 m/s, 18 m/s and 26 m/s, the
Models at 16m/s height increases by 31m, 9 m and 60 m respectively)

775

Authorized licensed use limited to: GE GLOBAL RESEARCH. Downloaded on June 19, 2009 at 08:26 from IEEE Xplore. Restrictions apply.
 4.1.2 Closed loop Response due to variation in
Altitude
The controller for Sarika-2 is designed based
on the mathematical model derived at straight and level
flight condition at an altitude of 1000m above the sea
level. However, the altitude at which the controller
operates might be very much different from that at
which it is designed. In order to analyze the
degradation in performance with altitude, the aircraft is
trimmed for level flight at altitudes of 1000, 1100 and
1200 m above sea level and its closed loop response
are simulated. Fig. 5 a, b and c shows the response of Figure. 5b Closed Loop Time Responses at different Altitude
the closed loop system after trimming the aircraft at
20m/s for a doublet command input. A pulse of -0.1 ms 5. Conclusion
(corresponding to elevator deflection of 3.84º) and This paper describes design and validation of robust
duration 2 seconds is applied at 0 seconds followed by fixed order H2 controller for micro air vehicle named
a pulse of 0.1 milliseconds again of duration 2 seconds. Sarika2. The controller performance is validated using
However, this set of simulation does not include sensor linear as well as nonlinear simulation models. Results
noise and gust disturbances. Responses initiated from were proved that designed robust stability
the trim speed of 20 m/s, shows that there is no much augmentation system is capable of providing the
variation in responses due to the altitude variations. desired closed loop requirements.
Thus, there will not be any performance degradation 6. References
due to height variations in real flight. Hence, the [1] Grasmeyer J. M., and Keennon M. T., Development of the
controller is robust against the variations in altitude. Black Widow Micro Air Vehicle, AIAA Paper 2001-0127,
Published by the American Institute of Aeronautics and
Astronautics, Inc.
[2] McMicheal J.M., and Francis M.S, Micro-Air Vehicles -
Toward a New Dimension in Flight World Wide Web,
http://www.darpa.mil/tto/mav/mav_auvsi.html 3/21/2003.
[3] Srinivasa Rao B.R., Surendranath V., and Prasanna H.R.S,
Wind tunnel test results of SARIKA Airplane, Report No:
AE/WT/IRR/16, Department of Aerospace Engineering, Indian
Institute of .Science, Bangalore, 2001.
[4] Bhat M. S., Meenakshi M., Govindaraju S. P., Surendranath
V., and Jamadagni. H. S., Design and Development of Flight
Control and Instrumentation for Micro Air Vehicle, Final
Technical Report No. ARDB/MSB/TR-99-1050, Department
of Aerospace Engineering, Indian Institute of .Science,
Bangalore, November 2002.
[5] Green M., and Limebeer, D. J. N., Linear Robust Control,
Prentice Hall, Englewood Cliffs, NJ, 1995.
[6] Liu Y., and Anderson., B.D.O., Controller Reduction:
Concepts and Approaches, IEEE Transactions on Automatic
Control, Vol. 34, No.8, pp. 802-812, August1989.
[7] Jan Roskam. Flight Dynamics part 4. Roskam Aviation and
Engineering Corporation Box 274 Ottawa Kansas 66067 USA
1971:5.1-6.118.
[8] Jan Roskam. Airplane design part 6: Preliminary calculation
of aerodynamic thrust and power characteristics. Roskam
Aviation and Engineering Corporation Box 274 Ottawa Kansas
66067 USA 1990: 213- 490.
[9] Pashilkar A.A., FAST User’s Manual Version 0.1, NAL PD
FC 0314, Flight Mechanics & Control Division, National
Aerospace Laboratories, September 2003.
[10] Meenakshi .M, Design and Real Time Validation of Discrete
Fixed Order Robust H2 Controller for Micro Air Vehicle, PhD
Figure. 5. Closed Loop Time Responses at different Altitude Thesis, Dept. of Aerospace Engineering, I.I.Sc Bangalore, July
2005

776

Authorized licensed use limited to: GE GLOBAL RESEARCH. Downloaded on June 19, 2009 at 08:26 from IEEE Xplore. Restrictions apply.