www.ijecs.

in
International Journal Of Engineering And Computer Science ISSN: 2319-7242
Volume 5 Issue 11 Nov. 2016, Page No. 19198-19211

Design and Modeling of Controllers for Aircraft Pitch Contol Movement

Ajai Kumar Singh And Rahul Dahiya*,
Department of Electrical Engineering, DCRUST Murthal, Sonepat, Haryana-131039,India.

ABSTRACT
In the present paper intelligent controllers design methodology is presented to examine their overall
performance primarily based on time response specification for controlling Pitch movement of aircraft. The
aircraft pitch control system is designed for Linear and Non-linear models and the effect of various
nonlinearities is studied based on the overall performance of various controllers in Flight control system
(FCS). The controllers are designed based on linearized model of aircraft so as to simplify the design
process, with the idea of dynamic modeling of Servomotor which is used to control the movement of
elevators. Elevators are control surfaces which are used to control longitudinal motion of aircraft by
deflecting the flow of air and creating differential pressure on the wings. Aircraft pitch control movement is
categorized under the longitudinal motion. The movement of these control surfaces is performed through
Servomotors, for which mathematical modeling is provided. A baseline reference controller is used for
determining the performance of other controllers and making a detail comparison. So while developing the
other controllers we took Fuzzy Logic Controller as reference as carry forwarded the original cased study. A
new hybrid combination of Fuzzy-PID with fuzzy compensation and ANN-PID Controller has been
developed for controlling the pitch movement of aircraft for this particular model. A quantitative analysis
of these controllers has been carried out in MATLAB Simulink© environment. It is concluded that
combination of ANN-PID controller provides best results for Linear as well as Non-linear aircraft pitch
control models.

Keywords- ANN, Pitch, Elevators, Aircraft, Flight during which aircraft changes its transition from
Control System, FLC. one state to another. There are three basic
control movements of aircraft i.e., Pitch, Roll, and
1. INTRODUCTION Yaw. The pitch movement of aircraft is
categorized under longitudinal stability whereas
There are three basic control movements of
roll and yaw are categorized under lateral
Aircraft i.e., Pitch, Roll and Yaw . One of the
stability.
important parameters out of these is Pitch
control movement. In the present work we have A set of control surfaces known as elevator are
considered this problem as valuable approach used for controlling aircraft pitch movement.
and carried out work by applying intelligent Elevators are movable control surfaces located at
control techniques for controlling the longitudinal the back of fixed wing aircraft and hinged to the
motion of Aircraft. Aircraft Pitch control trailing edge of horizontal stablizer, running
movement is a critical phase during takeoff phase parallel to the main wings that cause this rotation
as well as during steady flight. This parameter is of aircraft. The elevators cause the aircraft to
of a concern as Aircraft can go into a stall climb and descend and also to obtain sufficient
condition if Pitch control of Aircraft is not lift from the wings to keep the aircraft in level
calibrated properly. This is an important stage flight at various speeds[1].

Ajai Kumar Singh, IJECS Volume 05 Issue 11 Nov., 2016 Page No.19198-19211 Page 19198
 DOI: 10.18535/ijecs/v5i11.88

The elevators are movable control surfaces which or by optimized methods like genetic algorithms
can be moved up or down. If the elevator is to get better control and enhancement in the
rotated up, it decreases the lift force on the tail performance [3]. The third technique is based on
causing the tail to lower and the nose to rise. If Fuzzy-PID Controller used to control Aircraft
the elevator is rotated downward, it increases the Pitch. Fuzzy logic is used to tune each parameter
lift force on the tail causing it to raise and the of Proportional-integral-derivative (PID)
nose to lower. Lowering the aircraft’s nose controller by selecting appropriate fuzzy rules
increases forward speed, and raising the nose through simulation in Matlab and Simulink. The
decreases the forward speed [1]. The following developed model describes the aircraft motion
Fig.1 shows basic control surfaces of Aircraft. and its real dynamics accurately compared to
other proposed models that allow studying and
evaluating some pitch controllers beginning from
traditional controller systems, fuzzy controller
systems, and hybrid controller systems. The
fuzzy-PID controller offered the best response by
mixing the features of both the fuzzy and the PID
controllers, and the ability to adapt to the fuzzy
rules. [4]. The fourth kind of control mechanism is
derived by using Lyapunov Theory. According to
this technique, it is possible to ensure, under a
certain condition, the asymptotical stability of the
helicopter [5]. The fifth control strategy used in
literature is based on Linear Quadratic Regulator
Fig.1.1 Basic control surfaces of Fixed wing (LQR). As utilized by, the main advantage of this
Aircraft technique is that the optimal input signal turns
out to be obtainable from full state feedback (by
A lot of work has already been initiated in this
solving the Ricatti equation). But the analytical
field. Some of the already applied techniques are
solution to the Ricatti equation is difficult to
mentioned here: The first kind of control
compute [6]. The sixth control method used in
technique used is based on Fuzzy active
the literature is of comparative assessment of
disturbance rejection control. In this paper, an
LQR and Fuzzy Logic controller for pitch control of
active disturbance rejection control (ADRC)
aircraft [7]. The seventh control technique used is
strategy based on fuzzy control is proposed, so as
based on Adaptive control of the aircraft pitch
to realize the high performance
angular motion by using the dynamic inversion
of aircraft pitch control. The nonlinear active
principle. The obtained
disturbance rejection controller is designed to
adaptive control structures consist of parametric
improve the ability of anti-interference mean
estimators, dynamic compensators and command
while, fuzzy control is adopted to adjust the ADRC
filters. The estimated state and the estimated
parameters online, which
output are calculated with respect to the
makes control performance better [2]. The
estimated vector of the aircraft parameters. The
second technique used is based on GA tuned LQR
obtained control system is particularized for the
and PID controller for aircraft pitch control.
adaptive control of the pitch angular rate and the
According to this technique Genetic Algorithm
pitch angle, respectively, in the case of an F-
(GA) is used for tuning the parameter LQR and
15 aircraft whose flight may be affected or not by
PID controllers. The controller design begins with
wind shears [8].
the appropriate mathematical model to account
the longitudinal motion of an aircraft. LQR and In the present work we have considered a
PID controllers are associated with their standard linearized model of a fixed wing aircraft
parameters which can either be tuned manually for controlling its pitch movement. Although

Ajai Kumar Singh, IJECS Volume 05 Issue 11 Nov., 2016 Page No.19198-19211 Page 19199
 DOI: 10.18535/ijecs/v5i11.88

there are different aircraft models available, but from pilot is given in case of control
in the present work we have taken the reference mode, servo unit is designed to move the
of original case study which shows comparative elevators [9].
assessment of P, I, D, PI, PD, PID, and FLC, and
shows the effect of various non-linearities on the
aircraft system behavior, and how it is able to
respond to non-linearities. Now continuing this
work further, we developed different control
techniques namely ANN. We have also developed
a combination of ANN-PID controller and PID
CONTROLLER WITH Fuzzy compensation
technique. All the controllers are developed in
MATLAB Simulink environment and comparative
assessment of each controller is presented and
discussed.

Fig.2.1 Pitch channel operation in altitude hold mode and

2. MATHEMATICAL MODELING FOR synchronization mode block diagram

AIRCRAFT PICH CONTROL

MOVEMENT
In this section of the paper, a brief description for
modeling of aircraft pitch longitudinal movement
is discussed. The following sub section discusses
the components used in pitch control movement.

COMPONENTS USED IN PITCH CONTROL

There are four set of components which are used

for pitch control. These are:

a. Vertical gyro: The main function of

vertical gyro is to generate electrical
signals proportional to the aircraft’s pitch
attitude command.
b. Rate gyro: Whereas the rate gyro
function is to generate electrical signals Fig.2.2 Pitch channel operation in control mode block
proportional to the aircraft’s pitch rate diagram
command.
2.1 MODELING OF COMPONENT USED IN
c. Altitude controller: The altitude
controller is designed so as to give AIRCRAFT PITCH CONTROL
altitude hold signal when altitude hold The component which is modeled in the present
mode is selected which is shown in paper is the servo motor, for which following
Fig.2.1 transfer function is derived. Since aircraft pitch
d. Servo unit: The servo unit is considered control movement is controlled using elevators,
to be an important part as it is used to which deflect the air flow, causing aircraft to
control the elevator control surfaces. climb or descend by creating pressure on wings.
When the error signal is applied in case of And the movement of these elevators is
stabilization mode and control signal controlled using servo motor as shown in fig.

Ajai Kumar Singh, IJECS Volume 05 Issue 11 Nov., 2016 Page No.19198-19211 Page 19200
 DOI: 10.18535/ijecs/v5i11.88

Also, since we have considered generalized Transfer function of servomotor

aircraft model, so the prime focus is on the servo = = (2.6)
unit. ( ) [ ( ) ]

Considering the values

Transfer function of pitch control system
component

Transfer function of servo amplifier = ( )

Transfer function of gears = 1/15 = 0.66

Transfer function of pitch gyro =

Transfer function of pitch rate gyro =

To get transfer function of the servomotor, we , ,

have to model its components.
Transfer function of servomotor can be written as
Rotor circuit
( )
(2.1) Further in the original research the PID controller is
designed for aircraft pitch control using root locus
Rotor EMF technique for compensate and uncompensated
system. The following transfer function is obtained
(2.2) [9].

Mechanical torque G(s) =

(2.3) 2.2 SIMULINK DIAGRAM FOR PITCH CONTROL.

Equations governing mechanical port In the present paper, we have considered two
= ( ) (2.4) models for aircraft pitch control, one as the linear
model and another one with non-linear model.
(2.5) Following are the three nonlinearities which are
introduced in the pitch control system. These
non-linearities are either deliberately introduced
or they are present in the system itself. These are
as follows:

1. Dead Zone Non-Linearity

2. Saturation Non-Linearity
3. Backlash Non-Linearity

The effect of Dead-zone nonlinearity is

considered in pilots control column, Saturation
nonlinearity is considered in servo amplifier and
the effect of backlash nonlinearity is considered
in pilot’s control column. *9]

The following Fig.2.4 shows the pitch control

Simulink model for aircraft for linear and
Fig.2.3Schematic of Servomotor
nonlinear system.

Ajai Kumar Singh, IJECS Volume 05 Issue 11 Nov., 2016 Page No.19198-19211 Page 19201
 DOI: 10.18535/ijecs/v5i11.88

In our present work we have carried forward the

work and developed new control techniques and
designed the membership function required for
our process.

A control system is generally required to meet

following time response specifications: steady-
state error, damping factor and settling time. If
rise time is also taken into consideration then it
would consistent in the specification of settling
time as both of these depend upon steady-state
error and damping factor. By considering
proportional, integral and derivative error
compensation the controller which meets all
these specification is PID controller, which is
Fig.2.4 Aircraft pitch control for linear and non- discussed below.
linear model

3.1.1 PID CONTROLLER WITH FUZZY

3. METHODOLOGY COMPENSATION
3.1 Design of Controllers

To control a process variable some control The Fuzzy-PID control structure used in the
strategies are essential. A control strategy present work includes PID and Fuzzy logic
consists of two aspects: controller, both of them arrange in cascade.
(1) Control configuration and Output from the process is again feed back to the
(2) Controller. fuzzy and the corresponding output of fuzzy to
Control configuration can be further categorized the PID controller again. This technique
as:/feedback control configuration, feed forward mentioned here is also known as compensation
control configuration, cascade control technique. In this technique PID controller is
configuration etc. The second aspect of a control tuned online by using compensator formula and
strategy is the controller, which gives the desired gains are obtained. The membership
actuating signal to an actuator on the basis of functions developed for aircraft pitch control are
computed error, which is calculated from the set also fine tuned manually so as to keep the error
point and measured value of process variable. within tolerance limit. The membership functions
utilized here are triangular membership
In the present paper, we have designed two new functions. Now let’s discuss both these controller
controllers: PID Controller with Fuzzy individually.
Compensation and ANN and combination of
ANN-PID Controller which are discussed in PID CONTROLLER
following subsections. In the original case study
PID and FLC were developed and comparison was PID controller is widely used controller in process
made for linear and nonlinear system. In the control industries. For determining the values Kp,
original work a comparative assessment was Ki, and Kd the controller needs to be tuned.
done between different controllers and was Different tuning techniques are available for
concluded that FLC was able to handle determining the values of the gains. The
nonlinearities better. controller input is the error between the desired
output and the actual output. This error is
manipulated by the controller to produce a

Ajai Kumar Singh, IJECS Volume 05 Issue 11 Nov., 2016 Page No.19198-19211 Page 19202
 DOI: 10.18535/ijecs/v5i11.88

command signal for the process according to the output. The block diagram of FLC is shown in
relationship as given below. Fig.3.2.

( ) ( ) ∫ ( ) ( )
(3.1)

The above equation represents control law

known as proportional-integral-derivative (PID).
The gains of Kp, Ki, Kd, are suitably chosen by a
design process called PID tuning in order to
achieve desired transient response, as well as
zero steady-state error for given desired output
function. Due to excellent properties, PID
controller is commonly used in closed-loop
devices, especially in SISO plants. A block
diagram of PID controller is shown in Fig. 5. In
some flight control applications, the classical PID Fig.3.2 FLC block diagram
control may not offer the most efficient choice of
feedback control, especially when multiple inputs
and outputs are involved [10].
i. Fuzzification
( ) The fuzzification operation, F can be defined as
follows:
(3.2)
F: Ui Ui*

The fuzzification transforms ui to a fuzzy set ̃ ,

defined on the Ui, where
̃

ii. Rule Base

Fig. 3.1 Block diagram of PID controller A linguistic variable can be characterized by (1)
name of the variable, (2) its linguistic fuzzy sets,
(3) universe of discourse, (4) “syntactic rule” of
fuzzy sets, and (5) “Semantic rule” of fuzzy sets.
FUZZY LOGIC BASE CONTROLLER
Fuzzy if-then rule
The FLC structure consists of four main building In the simplest form, “fuzzy if-then rule” can be
blocks: (1) The fuzzifier that maps crisp input represented as:
either in direct form or normalized form i.e.
input ranges between [-1 , 1], to corresponding If x1 is A and x2 is B then y is C,
type-1 fuzzy set, (2) The “rule base” consists of
set of rules that depicts the knowledge of Where A, B and C are linguistic values defined by
designer about actions to be taken by the fuzzy sets on the universe of discourse X1, X2 and
controller, (3) The fuzzy inference system (FIS), Y, respectively. The part between ‘if and then’ is
interpret the type-1 fuzzy input set to type-1 called “antecedent” and the part after ‘then’ is
fuzzy output set according to rules provided in called “consequent”.
rule base, and in the last, (4) The defuzzifier iii. Fuzzy Inference System
convert type-1 fuzzy output set of FIS to a crisp

Ajai Kumar Singh, IJECS Volume 05 Issue 11 Nov., 2016 Page No.19198-19211 Page 19203
 DOI: 10.18535/ijecs/v5i11.88

There are two types of FIS that are widely used in of the general defuzzification methods are
applications: (1) Mamdani FIS [11] and Takagi- explained in brief as follows.
Sugeno-Kang (TSK) FIS [12]. The differences
between these two FIS lie in the consequent part Centroid of area
of their fuzzy rule, and thus they have different
defuzzification accordingly. Using centroid of area method the crisp output,
Mamdani FIS ZCOA is given as:
In Mamdani FIS the consequent part is also ∫ ( )
represented by a fuzzy set. Min-Max composition (3.1)
∫ ( )
of Mamdani FIS is generally used. Fig.5 shows
how a two rule Mamdani FIS compute the overall Where A is the output fuzzy set to be defuzzified
output fuzzy set defined in the universe of discourse z, ( ) is
the aggregated output of MFs.
when subject to crisp input x1 and x2 , using Min-
Max composition of Mamdani FIS [11]. Bisector of area

μ μ M μ
i The crisp output, ZBOA satisfies the following
n equation:

μ μ μ ∫ ( ) ∫ ( )
(3.2)

x x μ M Where α= min {z∣z є Z} and β= max {z ∣z є Z}.

1 2 a
x Mean of maximum

Fig.3.3 The Mamdani FIS using Min and Max In this method, ZMOM is the average of the
operation maximizing z at which MF reach a maximum μ*,
TSK FIS which can be mathematically represented as:
In TSK FIS, the consequent part is represented by
a conventional a polynomial function. A typical
TSK fuzzy rule has the form: ∫
∫
IF x is A and y is B then z=f(x,y);
(3.3)
Where A and B are input fuzzy sets in the
antecedent and f(x,y) is a polynomial of input Where z' = {z∣μA(z) = μ*}.
variable x and y in the consequent. However,
f(x,y) can be any function, until the output Smallest of maximum
defined by it is within the fuzzy region specified
by the antecedent of the rule [13].
In this method, the crisp output ZSOM is the
iv. Defuzzification minimum of the maximizing Z.

Largest of maximum
Defuzzification is a process to extract crisp output
from the type-1 fuzzy output set, provided by FIS.
There are various defuzzification methods, some

Ajai Kumar Singh, IJECS Volume 05 Issue 11 Nov., 2016 Page No.19198-19211 Page 19204
 DOI: 10.18535/ijecs/v5i11.88

In this method, the crisp output ZLOM is the NL NM NS ZE PS

maximum of the maximizing Z. NS
NM NS ZE PS PM
ZE
NS ZE PS PM PL
PS
ZE PS PM PL PVL
PM

PID controllers generally do not work very well

for nonlinear systems, higher order and time-
delayed linear systems, and particularly complex
Fig.3.4 Membership functions for input (error) and vague systems that have no precise
mathematical models. To overcome these
difficulties, various type of modified conventional
PID controllers such auto-tuning and adaptive PID
controllers.
Fig.8 shows the structure of Fuzzy-PID controller
which has been developed for the aircraft pitch
and roll control model. This model of Fuzzy-PID
controller developed here is using fuzzy
compensation technique. The purpose of
compensation is to achieve the desired
Fig.3.5 Membership functions for input (rate) performance of the system. By using
compensation we feed the output of the process
back to FLC as input, till the process becomes
stable.

Following values are obtained after tuning the

controller in Matlab:
Proportional (P) = 5.19667069406397

Fig. 3.6 Membership functions for output (control Integral (I) =0.960069009932765
action) Derivative (D) =6.98307455878522

Filter Coefficient (N) = 144.222740324956

Table 1 Rule Table for FLC 3.1.2 ARTIFICIAL NEURAL NETWORK

Simulated Neural Network is a control instrument

e/de NM NS ZE PS PM
utilized as a part of control hypothesis and is
extremely successful in flight control framework
NVL NL NM NS ZE planning.
NM

Ajai Kumar Singh, IJECS Volume 05 Issue 11 Nov., 2016 Page No.19198-19211 Page 19205
 DOI: 10.18535/ijecs/v5i11.88

In the present work, we are considering an O=Γ| |

elementary feed forward architecture of one Where W is the weight matrix, also called the
neuron receiving two inputs. Its output and input connection matrix: When the accompanying
vector are, respectively system is executed the neural system begins
preparing the information and executes a solitary
[ ] ANN piece. This neural system square is
[ ] supplanted in the pitch control SIMULINK
Weight connects the i’th neuron with the j’th obstructs with alternate controllers and the
input. The double subscript convention used for relating results are being noted. The
weights is such that the first and second subscript accompanying figure demonstrates the SIMULINK
denote the index of the destination and source piece outline for aircraft pitch control with ANN
nodes, respectively. The activation value for the [14].
i’th neuron as

=∑ for i=1,2,…,m 3.1.3 ANN-PID CONTROLLER

The following nonlinear transformation eq. The PID controller has a simple structure and thus
involving the activation function ( ), for i= the performance of the system can be affected in
1,2,…, m, completes the processing of x. The terms of time response specification. Also, the
transformation, performed by each of the m conventional PID control method has poor
neurons in the network, is a strongly nonlinear adaptation ability, and it is difficult to find the
mapping expressed as optimal PID gains. Thus a sophisticated control
scheme is required for an enhanced control
( ) for i=1,2,…,m
performance. The PID and neural network are
Where weight vector contains weights leading combined here to get optimum results. The
towards the i’th output node and is defined as placement of controllers is arranged in series as
follows shown in the Simulink block diagram 4.5. A
Proportional-Derivative controller is also placed
[ ] in the forward loop. Both the PID and PD
controller are tuned in Matlab and optimal gains
are obtained for best performance. The values of
and are obtained by tuning the
controller in Matlab.The values of , and
are mentioned below.
Following values are obtained for PID and PD
controller after tuning them.

Values of PID controller for linear system

Proportional (P) =39.4391254092514

Integral (I) =0.269108120085848

Fig.3.7 Single layer feed forward network Derivative (D) =0.429017553956734

Introducing the nonlinear matrix operator Γ. The Filter Coefficient (N) = 228.514492280641
mapping of input space x to output space o
Values for PD controller are mentioned below:
implemented by the network can be expressed as
follows Proportional (P) = 14.5465527331703

Ajai Kumar Singh, IJECS Volume 05 Issue 11 Nov., 2016 Page No.19198-19211 Page 19206
 DOI: 10.18535/ijecs/v5i11.88

Derivative (D) = 1.45359581889392

Filter Coefficient (N) = 68.3898684841076

Values of PID controller for Non-linear system

Proportional (P) =45.5970735385281

Integral (I) =0.307431266372757

Derivative (D) =0.498431066408624

Filter Coefficient (N) =262.369741876539

Values of PD Controller are mentioned below:

Proportional (P) = 19.204099695506

Derivative (D) = 2.37769924713268

Filter Coefficient (N) = 744.961689197853 4.3 Aircraft pitch control model using ANN for
Linear and Non-linear systems

4. MATLAB Simulink models for Aircraft Pitch

control

4.1 Aircraft pitch control model using PID

Controller for linear and non-linear systems

4.2 Aircraft pitch control model using FLC for 4.4 Aircraft pitch control model using PID
linear and non-linear systems controller with Fuzzy compensation for Linear
and Non-linear systems

Ajai Kumar Singh, IJECS Volume 05 Issue 11 Nov., 2016 Page No.19198-19211 Page 19207
 DOI: 10.18535/ijecs/v5i11.88

5. SIMULATION RESULTS AND DISCUSSION

In this part of paper simulation, results for

aircraft pitch control using different control
techniques is presented. Following responses are
obtained for SIMULINK models of Aircraft pitch
control movement discussed in the previous
section of paper. Using MATLAB SIMULINK model
we can depict the behavior of the system as it
would do in real conditions. MATLAB provides a
real-time observation using PID, FUZZY, ANN,
ANN-PID and FUZZY-PID controller to analyze the
performance of the aircraft. All the SIMULINK
models are created using MATLAB 2014a.

Fig. 5.1 and 5.2 shows step response of Aircraft

for Pitch control movement for Linear and Non-
linear system respectively. These responses show
a comparison of five different controllers namely
PID, FUZZY, FUZZY-PID, ANN, ANN-PID Controllers
developed for controlling Pitch movement of
Aircraft. The response of these controllers is
measured in terms of time response specification
i.e., Settling Time, Rise Time, Overshoot and
Steady State error.

For Linear and Non-linear Aircraft Pitch control

model, the simulation is performed for 150 sec.
From the fig. 5.1 and Fig. 5.2 following conclusion
is drawn:
4.5 Aircraft pitch control using ANN-PID
controller for Linear and Non-linear systems All the controllers achieve stability
instantaneously following the input. Though
there is steady-state error in performance of all
the controllers. This steady state error may be
neglected, since we have considered a
generalized model of Aircraft for pitch control
movement. Overshoot may be observed in three
controllers namely PID, FUZZY and ANN, which is
not of a concern and may be neglected as well.
For Non-linear Aircraft Pitch control model only
one controller shows slight deviation after 130
sec of simulation, which could also be neglected
as it is very minor, and the controller still follows
the input. No overshoot is observed from any
controller in Non-linear Aircraft Pitch control
model and all the controllers get’s settled
instantaneously.

Ajai Kumar Singh, IJECS Volume 05 Issue 11 Nov., 2016 Page No.19198-19211 Page 19208
 DOI: 10.18535/ijecs/v5i11.88

The following conclusions are made from the

table 5.1 and 5.2 that combination of an ANN-PIDE
controller is able to handle the Pitch controlL
E
movement of Aircraft better when detailV
comparison is done with other controllers inA
terms of time response specification. It has alsoT
been observed that PID controller cannot handleOR
the process when different non linearities are
introduced, making the system unstable. Fig. 5.2D
showing the response of PID controller is plottedE
F
when Backlash Non-linearity is not taken intoL
consideration. E
C
T
I
O
E
N
L
E Fig.5.2 Step response of Aircraft Pitch Control for Non-
V
linear System
A
T
O
R Table 5.1 Comparison of Different Controllers
for Aircraft Pitch Control using Linear model
D
E
PARAMET
F CONTROLLERS
ERS
L
E FUZZ
FUZZ ANN
C PID U- ANN
T
Y -PID
PID
I
O
N Settling 8.3 6.54 5.42 6.45 3.15
Time sec sec sec sec sec

Fig.5.1 Step response of Aircraft Pitch control for Linear Rise Time 2.9 2.36 1.5 2.52 0.55
System sec sec sec sec 2 sec
Overshot 2.93 5.58 5.55
14% 4%
% % %
Steady
State
0 0.35 0.38 0.36 0.48
error

Table 5.2 Comparison of Different Controllers

for Aircraft Pitch Control using Non-Linear model
PARAMET PID FUZZ FUZZ ANN ANN-
ERS Y Y- PID
PID
Settling
Time 26. 22.9 5.42 22.3 2.51
4 Sec sec sec sec
sec

Ajai Kumar Singh, IJECS Volume 05 Issue 11 Nov., 2016 Page No.19198-19211 Page 19209
 DOI: 10.18535/ijecs/v5i11.88

Rise Time For aircraft pitch control movement for linear and
16. 12. 1.24 11.8 0.277 non-linear model combination of ANN-PID
52 sec sec sec sec
sec
controller is able to handle the system better as
Overshoot compared to other controllers providing best
0% 0% 2.93 2.94 1.93 transient response ( in terms of time response
% % % specification). For obtaining the desired results
Steady we have placed a PD (Proportional Derivative)
State error 0.6 0.15 0.15 0.68 0.5 controller in the loop so as to compensate for any
6
distortions and getting the desired output. The
combination of ANN-PID controller is more
robust compared to other controllers.

From the Table 6.2 it is also concluded that for

6. CONCLUSION non-linear model individual controller cannot
handle the non-linearities effectively, and thus
In this paper, Pitch control of aircraft has been having larger settling time. And when the hybrid
studied. The various controller’s viz. PID, FLC, combination is implemented then the time
ANN, and a combination of PID with Fuzzy Logic response specification of the system improves
compensation, and ANN-PID based controllers drastically.
have been designed in MATLAB Simulink
environment. With the gained results the Some other intelligent and adaptive controllers
following comparisons have been accomplished such as Neuro-Fuzzy, Adaptive Fuzzy, ANN based
as shown in Table 6.1 and Table 6.2. The controllers can also be employed for the problem
performance of these controllers has been undertaken in this paper by considering detail
analyzed in terms of common performance design aspect of aircraft. Controllers used in this
criteria for Linear and Non-linear system report can be optimized using various
respectively for aircraft Pitch control. Optimization techniques such as Ant-colony
optimization, Genetic Algorithm, Particle Swarm
Further, it may me noted that detail design Optimization, Biogeography Based optimization
aspect of Aircraft Parameters i.e., Pitch control etc. In this paper, all the observation are made
movement is not taken into consideration. In the without taking into account the effect of
present work, we have considered a standard disturbances which occur in the environment
model for Aircraft Pitch control movement i.e., of acting on a body of Aircraft in the air, such as
the servomotor which is used to control the Hydrodynamic forces, radiation force, excitation
Elevator deflection which is used to control the force and Drag Force. Further, the techniques
pitch movement of Aircraft. Therefore all implemented in the present work can also be
controllers have marginal steady state error implemented for Yaw and Roll control
which may be neglected for the present aircraft movement of aircraft. The following Parameters
pitch control model. of Aircraft can also be designed using different
intelligent techniques by considering details
Hence, following conclusions are made from the
dynamics of the Aircraft.
results:
REFERENCES
PID controller is not able to handle the aircraft
pitch control movement for the non-linear [1] Airframe and Power plant Mechanics (AC 65-
system, especially when Backlash a non-linearity 15A)-Airframe Hand Book FAA.
is taken into consideration and the system is not
[2] D. We, H. Xiong, and J. Fu, “ Aircraft autopilot
able to linearize. The performance of PID pitch control based on Fuzzy Active
controller could only be evaluated for Linear Disturbance Rejection Control”, IEEE
System. Conference on Industrial Informatics-
Computing Technology, Intelligent
Ajai Kumar Singh, IJECS Volume 05 Issue 11 Nov., 2016 Page No.19198-19211 Page 19210
 DOI: 10.18535/ijecs/v5i11.88

Technology, Industrial Information

Integration (ICIICII), pp.144-147, 2015.
[13] S. Kamalasadan and A.A. Ghandakly, “A Neural Network
[3] Vishal and J. Orhi, “GA tuned LQR and PID Parallel adaptive Controller for
controller for aircraft pitch control”, IEEE
international conference on Power Fighter Aircraft Pitch Rate-Tracking”, IEEE Transactions on
Electronics (IICPE), pp.1-6, 2014. Instrumentation and Measurement, Vol.60, pp.258-267,
2011
[4] I.N. Ibrahim and M.A. Al Akkad, “Exploiting
an intelligent Fuzzy-PID system in nonlinear
aircraft pitch control”, IEEE International
Siberian Conference on Control and
Communication (SIBCON), pp.1-7, 2016.

[5] S. Kamalasadan and A.A. Ghandakly, “A

Neural Network Parallel adaptive Controller
for Fighter Aircraft Pitch Rate-Tracking”, IEEE
Transactions on Instrumentation and
Measurement, Vol.60, pp.258-267, 2011.
[6]
L. Chrif and Z.M. Kadda, “Aircraft Control
System Using LQG and LQR Controller with
Optimal Estimation-Kalman Filter Design”,
ELSEVIER 3rd International Symposium on
Aircraft Airworthiness, ISAA, Procedia Engg.,
Vol. 80, pp. 245-257, 2014.

[7] N. Wahid and M.F. Rahmat, “Pitch control

system using LQR and FLC”, IEEE Conference
on Industrial Electronics and Applications
(ISIEA), pp.389-394., 2010.

[8] D.K. Chaturvedi, “Modeling and Simulation of

Systems Using MATLAB and SIMULINK”, CRC
Press, pp. 435-473.

[9] A. Tewari, “Atmospheric and Space Flight

Dynamics”, Springer International Edition,
©2007 Birkhäuser Boston, pp. 471

[10]
E.H. Mamdani and S. Assilian, “An
experimental in linguistic synthesis with a
FLC”, International Journal of Man-Machine
Studies, Vol. 7(1). pp.1-13, 1975.

[11]
T. Takagi and M. Sugeno, “Fuzzy
identification of systems and its application
to modeling and control”, IEEE Transactions
on Systems Man and Cybernetics, Vol.15(1),
pp.116-132, 1985.

[12] T. Takagi and M. Sugeno, “Fuzzy

identification of systems and its application
to modeling and control, IEEE Transactions
on Systems Man and Cybernetics, Vol.15(1),
pp. 116-132, 1985.

Ajai Kumar Singh, IJECS Volume 05 Issue 11 Nov., 2016 Page No.19198-19211 Page 19211