IOP Conference Series:

Materials Science and

Engineering

PAPER • OPEN ACCESS You may also like

- A Review on Design Methods of Vertical
Sliding Mode Control of VTOL System take-off and landing UAV aircraft
Veneet Kumar, Rajkumar Sharma,
Shailesh Sharma et al.
To cite this article: Jerry Jacob and S Selva Kumar 2019 IOP Conf. Ser.: Mater. Sci. Eng. 561 012069
- Landing and take-off capabilities of
bioinspired aerial vehicles: a review
Ahmad Hammad and Sophie F Armanini

- PD control of VTOL aircraft trajectory

View the article online for updates and enhancements. tracking based on double-loop design
Wen-lan Wang, Xiong-huai Bai and Lei
Zheng

This content was downloaded from IP address 161.9.72.26 on 05/11/2024 at 12:10

ICMSMT 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 561 (2019) 012069 doi:10.1088/1757-899X/561/1/012069

Sliding Mode Control of VTOL System

Jerry Jacob, S Selva Kumar

Department of Electrical and Electronics Engineering,
Amrita School of Engineering, Coimbatore, Amrita VishwaVidyapeetham, India

Abstract. QUANSER QNET 2.0 Vertical Take Off and Landing (VTOL) system is a hardware device
setup developed by National Instruments to study flight dynamics and vertical take-off and landing
control. It has a DC Motor fan and adjustable counter weight. This is used exclusively for NI-ELVIS
platform and LabVIEW software. The LabVIEW software is used to identify the model of the system
by the method of best fit. The obtained second order model is tested for the open loop response using
MATLAB SIMULINK. The various control methods like Proporional+Integral+Velocity (PIV) and
Sliding Mode Control (SMC) are used to find the response of the system. It is found that the sliding
mode controller provides better results compared to the PIV controller. Along with this, the actuator
control is done using PI controller and it’s response along with the system is analyzed and compared
for both PIV and Sliding Mode Controller.

1. Introduction
Vertical Take-off and Landing aircrafts are those which possess the ability to hover, take –off and
land vertically. Classifications vary from fixed wing air-craft to helicopters. Some VTOL aircrafts
operates in other modes as well. These are CTOL (Conventional Take-Off and Landing), STOL
(Short Take-Off and Landing) and STOVL (Short Take-Off and Vertical Landing). In a flight or
airplane there are three types of motion which affect the stability and control. These are yaw, roll and
pitch. Roll is the rotation of the aircraft around the front-to-back axis. Pitch is the rotation of the
airplane from side-to-side axis. Whereas, the rotation around the vertical axis is known as yaw.

The salient features of the QUANSER QNET 2.0 VTOL system are high speed DC motor with a
safety guard, a propeller assembly, a digital tachometer, a current sensor, a PWM amplifier, optical
encoders to find the position of the arm and a protective cover. It has a built in Peripheral Component
Interconnect (PCI) connector which interfaces it to NI-ELVIS. This system is fully compatible with
LabVIEW software. The mass of the counter-weight is 258g. The resolution of the encoder is
0.176°/count. The peak current of the amplifier is 2.5A, while the continuous current is 0.5 A. The
output voltage of the amplifier is ±24V with 42% duty cycle. The QNET VTOL is designed for
indoor use only. The temperature which it can withstand is from 5° C to 40 ° C. It can bear up to an
altitude of 2000 m. The maximum relative humidity of 80% at 31°C which decreases linearly to 50%
relative humidity at 40°C. The mains supply voltage is ±10V of the nominal voltage. The maximum
transient overvoltage is 2500V. The pollution degree of the device is 2.

The PI controller is used for the actuator of the QUANSER QNET 2.0 VTOL system. The actuator
regulates the input current drawn by the system. Generally, the system uses the PIV controller. But the
sliding mode control is incorporated instead of PIV and analysis is done.

2. Literature Survey
Peter Martin et.al.[1] has done the basic modeling of the QUANSER QNET 2.0 VTOL system.
Through this, the second order transfer function of the VTOL system can be found out. Also Peter
Martin et.al.[2] had done the white box model of the system. Various parameters of the VTOL system
especially for the DC Motor fan and other components are present and the respective transfer function
can be derived. S Mondal and C Mahanta [3] had proposed a second order sliding mode controller in

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
ICMSMT 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 561 (2019) 012069 doi:10.1088/1757-899X/561/1/012069

order to stabilize the VTOL system. S Anoop and K.R.Sharma [4] had done simulation studies based
on the QUANSER QNET 2.0 VTOL system for Model Predictive Controller (MPC). The simulation
of the MPC is compared to that of the conventional PID controller. The authors [5] had done an
altitude control of a VTOL tail sitter Unmanned Aerial Vehicle (UAV). The pitch as well as the roll
control of the UAV is done. J Zhang et.al. [6], had designed a UAV with a platform structure for
multi-rotors and fixed wing rotors (FW) on conventional control surface. The mathematical model of
the UAV is discussed and control methodologies are implemented in this work. S. Niwa and I.
Sugiura [7] discussed about acontrol methodology for VTOL aircraft supported by ducted fan.
Various problems like stability, control and fuel efficiency were solved in this work. The authors [8]
had implemented a linear optimal design approach with multi-variable control laws in VTOL
aircrafts. A. Yang and Y. Wang [9] presented a new type of vertical take-off and landing (VTOL)
aircraft which alters the three elements of thrust. A plan for a non-linear segmentation is proposed and
the degree of correctness is found through numerous simulations. R.K.R.Anaet.al.[10] had discussed
about a new control design of the PIV controller. The compensator and controller are made using the
Eigen value assignment method. The authors [11] had designed a sliding mode controller for a boost
converter.This paper presents a model identification procedure for the VTOL system and compares
the system response for different controllers likePIV and SMC. Also, a PI controller is tuned for the
actuator and the overall system response is studied and compared for PIV and SMC incorporated in
the VTOL system.

3. Theoretical Background

Figure 1. Block Diagram of VTOL System

NI-ELVIS is known as NI Educational Laboratory Virtual Instrumentation Suite. It is an engineering

device used by students and professionals in academia. Figure 1.shows the interactions between NI-
ELVIS board and the VTOL system. The VTOL system consists of the PWM amplifier, fan, encoder
and the current sensor. The PWM amplifier is used to power on the high speed DC fan motor. As the
fan operates the system moves about the pivot. The encoder is used to find the position of the VTOL
system. A current sensor is used in order to measure the current taken by the PWM amplifier.

Figure 2.Cascade Control of a QUANSER QNET 2.0 VTOL System

2
ICMSMT 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 561 (2019) 012069 doi:10.1088/1757-899X/561/1/012069

Figure 2 depicts the cascade control of the QUANSER QNET 2.0 VTOL system. It is divided into
two subsystems: the voltage-current dynamics of the actuator i.e. the motor and the position dynamics
of the VTOL. The controller in the inner loop is designed to control or regulate the current of the high
speed fan DC motor. The reference current is generated from the outer loop using the controller
present in the outer loop. The outer loop is used to control the pitch of the system.

Figure 3.PI Control of the High Speed DC Motor

Figure 3 illustrates the PI control of the high speed DC motor i.e. the actuator. This is required when
the actuator of a system has slower dynamics. PI controller is used in order to regulate the flow of the
current to the entire system. This makes the actuator dynamics very small making the overall system
simpler.
Equation (1) shows the voltage-current relationship of the DC motor actuator,

dim
vm Rmim Lm (1)
dt
The voltage across the motor is vm, im current drawn by the motor with resistance Rmand inductance of
the motor is Lm.
Equation (2) shows the voltage-current relation in the transfer function form,

Vm
Im (2)
sLm Rm

The transfer function of the system has Vm i.e. voltage induced to the motor as the input to the actuator
and Im i.e. the output current to be induced to the system as output.

Figure 4.Free Body Diagram of VTOL System

Unlike the DC motor, the VTOL system can be characterized by a second order equation. The VTOL
system consists of propeller actuator, VTOL body and counter weight. Fig 4 shows the free body

3
ICMSMT 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 561 (2019) 012069 doi:10.1088/1757-899X/561/1/012069

diagram depicting these parts. Equation (3) gives the transfer function of the VTOL system [1] which
is derived from the free body diagram.

Kt
P( s) (3)
B K
J s2 s
J J

where, Kt-thrust current torque constant Nm/A

J -moment of inertia in kgm2
B -viscous damping in Nm/(rad/s)
K-stiffness of the VTOL system in Nm/rad

Referring [1] and [2] the parameters of the transfer function of the VTOL system were found as in
Table 1.
Table 1. Parameters of QUANSER QNET VTOL System
S.No Parameters Values
1. J 0.0045kgm2
2. B 0.002 Nm/(rad/s)
3. Kt 0.1562 Nm/A
4. Rm 40.2 Ω
5. Lm 2H
6. m1 258g
7. m2 127g
8. l1 72.5mm
9 l2 155mm
10. K 0.0218 Nm/rad

Using Table 1, the transfer function of the VTOL system can be found out using equation (3).
Equation (4) shows the transfer function of the system when the values mentioned in Table 1 are
assigned,
34.71
P( s ) 2
(4)
s 0.444s 4.84

Equation (5) shows the transfer function of actuator,

1
I m( s) (5)
2s 440.2

4. Controllers for VTOL system

PI controller is used for the control of the actuator or the high speed DC motor fan. It is a very popular
controller used in process control industries. PI controller is represented by Equation (6),

u K pe Ki e.dt (6)
Where,KP - Proportional Gain constant
Ki - Integral Gain constant
e - Error

The proportional gain, KP multiplies the error and it drives the output proportional to error. Integral
gain, Ki decreases the rise time, increases overshoot and settling time but it eliminates steady state
error.

4
ICMSMT 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 561 (2019) 012069 doi:10.1088/1757-899X/561/1/012069

PIV controller is used to control the pitch of the QUANSER QNET 2.0 VTOL system. It has two
components i.e. the feed forward control used to add auxiliary velocity and in some cases acceleration
signals are added to the servo loop in order to improve command tracking. These components play a
crucial role in rejecting disturbances. PIV controller is represented by Equation (7),

K i E ( s) K Ns
U ( s) K p E ( s) v (7)
s s N
where , KP - Proportional Gain constant
Ki - Integral Gain constant
Kv-Velocity constant
N-constant of the feed forward control
E(s) - Error in s domain

Figure 5. Sliding Surface or Manifold of the Sliding Mode Controller

Like the PIV controller, there also exist three parameters for sliding mode control. The parameters are
sliding mode gain, K, sliding mode slope, λ and sliding mode thickness, ϕ. These parameters are used
to control the system according to the sliding manifold or surface shown in Figure 5. As the sliding
mode thickness, ϕ increases, the chattering of the system response decreases. However, the steady
state error slightly increases as the thickness, ϕincreases.
Equation (8) gives the function of the sliding surface,

s e e  (8)
The sliding mode function, s is the sum of the products of slope, λ and error, e and derivative of error.
The derivative of the error is needed to do the tracking of a set point.
Equation (9) shows the switching function of the sliding mode controller,

sign( s) for s
Switching Function s (9)
for s

When |s| is greater thanϕ, the signum function of s is considered. Otherwise the function chosen is
s/ϕ. This function is taken to derive the output of the controller.
Equation (10) gives the output of the sliding mode controller,

Output (e K S
(e Switching Function) (10)

The output of the controller is the sum of the product of sliding mode gain, K and switching function
and derivative of error. This is introduced to the plant in order to reach the set point at the earliest.

5
ICMSMT 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 561 (2019) 012069 doi:10.1088/1757-899X/561/1/012069

5. Model Identification

Figure 6. Hardware Setup of VTOL System with Interface

The QUANSER QNET 2.0 VTOL is fixed to the NI-ELVIS hardware interface board. Fig 6 shows
the hardware setup of QUANSER QNET2.0 VTOL system. The NI-ELVIS board is connected to a
PC/Laptop via a USB cable. The software used for the model identification of the VTOL system is
LabVIEW 2015. LabVIEW is known as Laboratory Virtual Instrumentation Engineering Workbench.
The unique feature of LabVIEW compared to other software is that it makes use of graphical
programming or G-programming. This allows users to access the blocks required for a particular
function. LabVIEW consists of two windows: front panel and block diagram. Front panel is where the
user can see the results. On the other hand, block diagram is used to do the graphical coding.

Figure 7.Front Panel of VTOL System to find Balancing Current.

Initially the transfer function derived in equation (4) is used in Figure 7. A signal of amplitude 0 A,
frequency of 0.2 Hz and offset of 0.13 A is introduced into the system under use. It is observed that
the VTOL system remains still in the horizontal direction i.e. the pitch of the system is at 0°. Hence
the balancing current is found to be at 0.13 A.

Using this balancing current, the model identification of the VTOL system can be found out. A signal
of amplitude 0.02A, frequency of 0.2 Hz and offset of 0.13 A is used in this case to find out the
model. The transfer function from Equation (4) is used initially to find the best fit response. However,
it is observed that the best fit condition does not occur when equation (4) is used in the front panel.
Hence, it cannot be used.

6
ICMSMT 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 561 (2019) 012069 doi:10.1088/1757-899X/561/1/012069

Figure 8.Model with Best Fit

The estimated transfer function is found out in the front panel. On further iterations, the transfer
function of the system which has the best fit is found out. The response is depicted in Figure 8.
Equation (11) gives the best fit model of the system,

3.34285
P( s ) (11)
0.17284s 2 0.324603s 1

On further simplification, Equation (12) is found out,

19.36
P( s ) 2
(12)
s 1.8819s 5.7977

This transfer function in Equation (12) is used for simulations and analysis.

6. Controller Implementation
The simulation of VTOL system with the PIV controller is done and then compared with that of the
sliding mode controller. The Control System tuner app is used to tune the parameters Kp, Kiand Kv.
The feed forward control constant, N is considered as 25. Hence the transfer function for the velocity
parameter is shown in Equation (13),
25s
Tk ( s) (13)
s 225

Figure 9. Block Diagram of PIV Controller with VTOL System

7
ICMSMT 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 561 (2019) 012069 doi:10.1088/1757-899X/561/1/012069

Figure 9 shows the block diagram of the PIV controller with the QUANSER QNET 2.0 VTOL
system. A step signal is introduced to the system. The PIV controller consists of the PI block and a
gain block, Kv. This is used to control the outer loop of the system. The actuator is controlled by the PI
controller i.e. the inner loop. The parameters of the PIV controller are tuned using the Control System
tuner app, whereas the PI control of the actuator is done by using the auto tuning method which is
available in the MATLAB SIMULINK software. Using these tuning methods the parameters are Kp=
0.5666, Ki = 2.5065 and Kv= -0.4449 for the outer loop. While for the inner loop the parameters are
Kp= 100.8318 and Ki = 7507.5568.

Figure 10.Block Diagram of SMC with VTOL System

Figure 10 shows the sliding mode control of the VTOL system. Like the PIV control, a step function
is introduced to the system as the input. The MATLAB function is used to design the sliding mode
controller. Instead of the PIV controller, the sliding mode controller acts as the controller here for the
outer loop. The PI control of the actuator remains the same in this case. On using the trial and error
method for sliding mode controller for the outer loop, the parameters are found to be K=10, λ=1 and
ϕ=0.01, 0.04 or 0.08. While using the same tuning methodology for PI controller, the parameters are
found to be Kp=1 and Ki=2.

7. Results
The VTOL system is simulated using SIMULINK for two control methodology namely PIV control
and sliding mode control. The results were discussed for a step input variation to observe the change
in the response of the system. Various observations and corresponding inferences were discussed.

Figure 11.Actuator Response of PIV control for VTOL System

Figure 11 shows the actuator response of the system under PIV controller. Initially, the system at 0.02
seconds will reach a peak of 4 V. However, at approximately 0.2 seconds the voltage reaches a low of
-1 V. The system settles at 0.15 A with some chattering at approximately 0.6 seconds.

8
ICMSMT 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 561 (2019) 012069 doi:10.1088/1757-899X/561/1/012069

Figure 12.Response of VTOL System with PIV Controller

Figure 12 illustrates the response of the PIV control of the VTOL system. A step response of 0.5 is
given to the system. It is observed that the peak of the response is at 0.52° when the time is
approximately 0.3 seconds. After 0.3 seconds, uneven oscillations occur in the system. The system
settles at 12 seconds for 2% tolerance band. However, the system which is under the PIV control will
have some chattering effect as the system reaches the tolerance band.

Figure 13. Actuator Response of SMC for VTOL System

Figure 13 depicts the actuator response using PI controller where Kp=1 and Ki=2. However, it is
observed that the sliding mode thickness, ϕaffects the response of the actuator. Basically sliding mode
thickness is used to reduce the chattering effect present in the system. Table 2 shows the peak, settling
time and the voltage at which the system settles for different sliding mode thicknesses.

Table 2.Time Domain Specifications for Different Sliding Mode Thicknesses

Sliding mode Peak
S.No Peak time Settling time Settling point
thickness overshoot
1. 0.01 0.286 V 0.15s No settling time No settling point
2. 0.04 0.282 V 0.15s 4.4s 0.1495 A
3. 0.08 0.274 V 0.13s 4.4s 0.1494A

From Table 2, it is understood that when the sliding mode thickness, ϕ is less, the system fails to
settles and oscillates unevenly. But as the thickness, ϕ increases the actuator regulates itself to a value
and remains to be in settled state. It is also observed that there exists a settling time at which the
actuator settles as ϕ increases. The peak overshoot of the system increases as the sliding mode
thickness is decreased.

9
ICMSMT 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 561 (2019) 012069 doi:10.1088/1757-899X/561/1/012069

Figure 14. Response of VTOL System with SMC

Figure 14 depicts the response of the system using sliding mode control. A step function of 0.5° is
introduced to the VTOL system. The sliding mode thicknesses, ϕ used in this system are 0.01, 0.04
and 0.08 along with sliding mode gain, K=10 and sliding mode slope, λ=1. Generally, the system
settles at about 8.5 seconds. The chattering present in the actuator response can be viewed in system
response. Table 3 gives the peak overshoot, peak time, settling time and settling point in the system.

Table 3. Chattering and settling points for various sliding mode thicknesses
S.No Sliding mode thickness Chattering Settling point
1. 0.01 More Between 0.501° and 0.485°
2. 0.04 Less About 0.4995°
3. 0.08 Least About 0.4987°

Table 3 shows that when sliding mode thickness is less, chattering is more. When ϕ=0.01, the steady
state error may be difficult to find out due to the chattering present. As ϕ increases to 0.04, the steady
state error is -0.0005°. On increasing further to 0.08, the steady state error is found to be -0.0013°.
The settling time for the system with any sliding mode thickness, ϕ is 8 seconds. Therefore, it can be
inferred that the steady state error decreases with increase in the sliding mode thickness, ϕ.

8. Conclusion
The QUANSER QNET 2.0 VTOL system was studied and analyzed with different controllers. The
mathematical model of the system was identified using LabVIEW. The obtained mathematical model
is considered for the simulation studies The PIV and sliding mode controllers are implemented for the
VTOL system model and results were compared and analyzed for a step change in input. SMC
controller with different sliding mode thickness and its effect on the system response were studied. It
is inferred that the peak overshoot and settling time are found to be lesser for sliding mode controller
compared to PIV controller. From results obtained, it is understood that sliding mode control produces
better results compared to PIV controller for a VTOL system.

References

[1] Peter Martin et.al 2015 INSTRUCTOR WORKBOOK- QNET 2.0 VTOL Board for NI-
ELVIS (QUANSER).

[2] Peter Martin et.al 2015 USER MANUAL- QNET 2.0 VTOL Board for NI-ELVIS
,QUANSER3 18.

10
ICMSMT 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 561 (2019) 012069 doi:10.1088/1757-899X/561/1/012069

[3] S. Mondal and C. Mahanta 2013 Observer based sliding mode control strategy for
vertical take-off and landing (VTOL) aircraft system (Melbourne) ,pp.1-6.

[4] S. Anoop and K. R. Sharma 2018 Model predictive control: Simulation studies for the
implementation on vertical take-off and landing labprototype Procedia Computer Science,
vol. 143, pp. 663-670.

[5] S. Verling, B. Weibel, M. Boosfeld, K. Alexis, M. Burri and R. Siegwart 2016 Full
Attitude Control of a VTOL tailsitter UAV 2016 IEEE International Conference on
Robotics and Automation (ICRA)(Stockholm), pp. 3006-3012.

[6] J. Zhang, Z. Guo and L. Wu 2017 Research on control scheme of vertical take-off and
landing fixed-wing UAV 2017 2nd Asia-Pacific Conference on Intelligent Robot Systems
(ACIRS),(Wuhan) , pp. 200-204.

[7] S. Niwa, I. Sugiura 1987 An Automatic Flight Control System for VTOL Aircraft
Supported by Ducted Fans IFAC Proceedings Volumes, vol 20, issue 5, part 6, ,pp 145-
150.

[8] R. F. Stengel, J. R. Broussard and P. W. Berry 1976 Digital controllers for VTOL aircraft
1976 IEEE Conference on Decision and Control including the 15th Symposium on
Adaptive Processes, (Clearwater)(FL, USA) , pp. 1009-1016.

[9] A. Yang and Y. Wang 2017 A new VTOL aircraft2017 36th Chinese Control Conference
(CCC), (Dalian), pp. 6213-6218

[10] R. K. R. Ana, N. Choudhary, J. S. Lather and G. L. Pahuja 2014 PIV and lead
compensator design using Lambert W function for rotary motions of SRV02 plant 2014
IEEE 10th International Colloquium on Signal Processing and its Applications (Kuala
Lumpur) , pp. 266-270.

[11] V. Vijayakumar, R. Divya, and A. Vivek 2014 Sliding mode controlled quadratic boost
converter International Conference on Computation of Power, Energy, Information and
Communication (ICCPEIC)(Chennai).

11