sensors

Article
Design of Amphibious Vehicle for Unmanned
Mission in Water Quality Monitoring Using Internet
of Things
Balasubramanian Esakki 1 , Surendar Ganesan 1 , Silambarasan Mathiyazhagan 1 ,
Kanagachidambaresan Ramasubramanian 2 , Bhuvaneshwaran Gnanasekaran 3 ,
Byungrak Son 4, * , Su Woo Park 5 and Jae Sung Choi 6
1 Centre for Autonomous System Research, VelTech Rangarajan Dr. Sagunthala R&D Institute of Science and
Technology, Avadi, Chennai 600 062, India; balasubramaniane@veltech.edu.in (B.E.);
surendarg@veltechuniv.edu.in (S.G.); silambu.friends@gmail.com (S.M.)
2 Department of Computer Science Engineering, VelTech Rangarajan Dr Sagunthala R&D Institute of Science
and Technology, Avadi, Chennai 600 062, India; drgrkanagachidambaresan@veltech.edu.in
3 UCAL System Ltd., Chennai 600 018, India; gd.bhuvanesh1@gmail.com
4 Convergence Research Center for Wellness, DGIST, Daegu 42988, Korea
5 Rovitek Inc., 91, Damun-ro 61-gil, Jillyang-eup, Gyeongsan-si, Gyeongsangbuk-do 38479, Korea;
swpark@rovitek.com
6 Department of Computer Engineering, Sun Moon University, Asan-si 31460, Korea; jschoi@sunmoon.ac.kr
* Correspondence: brson@dgist.ac.kr; Tel.: +82-53-785-4772




Received: 13 September 2018; Accepted: 28 September 2018; Published: 3 October 2018


Abstract: Unmanned aerial vehicles (UAVs) have gained significant attention in recent times due
to their suitability for a wide variety of civil, military, and societal missions. Development of an
unmanned amphibious vehicle integrating the features of a multi-rotor UAV and a hovercraft is the
focus of the present study. Components and subsystems of the amphibious vehicle are developed
with due consideration for aerodynamic, structural, and environmental aspects. Finite element
analysis (FEA) on static thrust conditions and skirt pressure are performed to evaluate the strength of
the structure. For diverse wind conditions and angles of attack (AOA), computational fluid dynamic
(CFD) analysis is carried out to assess the effect of drag and suitable design modification is suggested.
A prototype is built with a 7 kg payload capacity and successfully tested for stable operations in flight
and water-borne modes. Internet of things (IoT) based water quality measurement is performed in a
typical lake and water quality is measured using pH, dissolved oxygen (DO), turbidity, and electrical
conductivity (EC) sensors. The developed vehicle is expected to meet functional requirements of
disaster missions catering to the water quality monitoring of large water bodies.

Keywords: amphibious UAV; hovercraft; FEA; CFD; prototype; water quality; sensors; Internet
of Things

1. Introduction
Unmanned aerial vehicles (UAVs) are categorized based on the performance characteristics of
their wing movement such as fixed-wing, rotary, and flapping-wing configurations [1]. Various
applications of UAVs include surveillance, traffic monitoring, active weapon engagement, wildlife
surveying, pollution monitoring, precision agriculture, etc. [2]. Authors of this paper contributed to
the development of UAVs for environmental monitoring [3], structural health monitoring [4], and also
constructed micro aerial vehicles [5–7]. Dedicated efforts on the development of amphibious vehicles
are scarce. Collins [8] described the importance of amphibious UAV in diverse applications and

Sensors 2018, 18, 3318; doi:10.3390/s18103318 www.mdpi.com/journal/sensors

Sensors 2018, 18, 3318 2 of 23

discussed their relevance in issues pertaining to control, communication, and airspace management.
Boxerbaum et al. [9] developed a robotic amphibious vehicle using biological concepts inspired by
animals to navigate in underwater and rough terrains. Yayla et al. [10] performed theoretical analysis
to investigate the performance characteristics such as rate-of-climb, turn radius, and maximum
velocity of an amphibious UAV. Pisanich and Morris [11] fabricated a sea plane conceptual model
of an amphibious UAV for a 4 kg payload. Autonomous flight missions were performed in air and
water as a proof of concept. Hasnan and Wahab [12] designed a UAV that can fly in air, and glide
along land and a water surface. Frejek and Nokleby [13] designed a four-paddle-wheel amphibious
vehicle with ultrasonic sensors to detect obstacles. The published data indicates that development
of amphibious UAVs for deployment in water quality assessment is not evident from the literature.
Measurement of water quality is usually performed with the aid of boats [14], which is labor-intensive
and costly. Remote sensing methods of water quality assessment is time-consuming and needs lot of
investment [15]. Traditionally, water body agencies are collecting the water samples manually and
in a periodic manner, which is cumbersome [16]. Few of the lakes, rivers, ponds, and reservoirs may
not have access to collect water samples with boats and they might be surrounded with shrubs and
bushes [17]. Also, in situ measurement of water quality at the designated locations of water bodies and
periodic measurement are of paramount interest to address the water quality [18]. An autonomous
underwater vehicle is deployed [19] and it suffers due to the global positioning system (GPS) denied
environment and controlling of those vehicles are difficult to collect water samples at the precise
water location. In order to overcome the aforementioned difficulties, a radical approach in collecting
water samples and performing in situ water quality assessment is necessary. UAV-based water quality
can be a relatively inexpensive, easy, and effective approach in accessing the large water bodies in
a short span of time. Landing of vehicles in dangerous water locations where human threat is of
major concern and inaccessible water bodies with boats are effortlessly accessed with the aid of UAVs.
Also, moving along the water body may be of peculiar interest to consume less power that necessitate
UAV with amphibious characteristics. For quality evaluation of large and inaccessible water bodies,
amphibious vehicles provide effective and rapid solutions. Development of a UAV that can land
and glide on the water’s surface while collecting water samples offers several challenges related to
materials, energy management, control systems, and on-board sensors. The present study integrates
features of a multi-rotor UAV with a hovercraft and this configuration for an amphibious UAV has not
been attempted yet. The vertical take-off and landing (VTOL) functionality is also integrated into the
system so that the resultant amphibious vehicle offers several functional advantages, such as energy
efficient movement on the water’s surface, eliminating large areas for landing and take-off besides
ensuring compatibility with a wide variety of payloads.

2. Evolving Conceptual Model

The conceptual model is formulated by integrating the multirotor and hovercraft configurations
wherein four co-axial propellers and motors that are attached to the frame act as an octo-rotor as shown
in Figure 1. The entire rotor assembly is supported by a hull made of high-density polyurethane foam,
and nylon impregnated with urethane is attached beneath for functioning as a skirt. Provision for
payloads, batteries, sensors, electronic accessories, flight controller board (FCB), water sampler with
robotic arm module, and water collection tanks are made in such a way that the center of gravity (CG)
of the vehicle is maintained for stable flight.
 • 2000 m of flight range
• Hovering and moving on water body
• Endurance of 40 min
Sensors•2018, 30
18, km/h
cruise speed
3318 3 of 23
• 2000 m range

Figure1.1.Conceptual
Figure Conceptualmodel
modelof
ofamphibious
amphibiousvehicle.
vehicle.

The co-axial
Based propellers
on these design are actuated during
requirements vertical take-off
and performance and alanding.
criteria, missionAfter
profilelanding
for theon a water
collection
surface, ◦
of waterany two co-axial
samples propellers
in remote are rotated
water body through
locations 90 using
is identified. Thea servo motor
typical and profile
mission these propellers
has flight
produce thrust for forward movement of the vehicle. Buoyancy of the vehicle
conditions of VTOL, hovering on air, landing on water bodies, propelling along the water is achieved through a
surface,
cushioning
and verticaleffect of theThe
landing. skirt produced
sequence ofusing
these amissions
duct fan.isAvaried
hoveraccording
gap of 2–5tommtheisoperational
maintained need.
between
the skirt of amphibious vehicle and the water surface. During hovering mode, all four co-axial motors
are powered
3. Design to create lift.
Process
In addition to these design requirements, the following performance specifications are considered
Design process of the amphibious vehicle capturing the functionalities corresponding to
for selection of aircraft components:
aerodynamics, structures, and compliance to performance criteria is presented in Figure 2.
• Airborne operation

• Flight endurance of 20 min

• Payload of 7 kg
• 40 km/h cruise speed
• 2000 m of flight range

• Hovering and moving on water body

• Endurance of 40 min
• 30 km/h cruise speed
• 2000 m range

Based on these design requirements and performance criteria, a mission profile for the collection
of water samples in remote water body locations is identified. The typical mission profile has flight
Sensors 2018, 18, 3318 4 of 23

conditions of VTOL, hovering on air, landing on water bodies, propelling along the water surface, and
vertical landing. The sequence of these missions is varied according to the operational need.

3. Design Process
Design process of the amphibious vehicle capturing the functionalities corresponding to
aerodynamics, structures, and compliance to performance criteria is presented in Figure 2.
Sensors 2018, 18, 3318 4 of 22

Performance Criteria for Amphibious Vehicle

Design Specifications

Conceptual Sketches

Design Calculations

Weight Estimation

Aerodyanmic Studies

Strutural Analysis

Fabrication and Testing

Figure 2. Design strategy of the amphibious vehicle.

3.1. Design
3.1. Design of
of Hovercraft
Hovercraft

AA hovercraft
hovercraft isisan
anair
aircushion
cushionvehicle
vehiclethat moves
that moves over multiple
over terrains
multiple including
terrains land,
including water,
land, and
water,
muddy
and muddysurfaces. The duct
surfaces. The fan
ductlocated at the center
fan located at the produces the necessary
center produces cushioning
the necessary effect through
cushioning effect
forcing air down and creating an air cushion between the skirt and the water
through forcing air down and creating an air cushion between the skirt and the water surface. surface. Inflation of
the skirt increases air pressure that acts at the base of the hull. Forward motion of
Inflation of the skirt increases air pressure that acts at the base of the hull. Forward motion of the the hovercraft is
achieved through
hovercraft propelling
is achieved through thepropelling
co-axial rotors. Since the
the co-axial skirtSince
rotors. is considered to considered
the skirt is be a sensitive part
to be a
of the hovercraft to achieve lift of the vehicle, selection of skirt material is an important
sensitive part of the hovercraft to achieve lift of the vehicle, selection of skirt material is an important aspect that
is discussed
aspect that isindiscussed
the following
in thesection.
following In order to In
section. design
orderato
hovercraft, various parameters
design a hovercraft, have to be
various parameters
determined. Table 1 presents the list of assumptions incorporated to perform
have to be determined. Table 1 presents the list of assumptions incorporated to perform design design calculations, and
in Table 2 hovercraft parameters are estimated for the flow analysis [20].
calculations, and in Table 2 hovercraft parameters are estimated for the flow analysis [20].

Table 1. Assumptions of the hovercraft parameters.

Serial No. Empirical Relation Limit

1 Length to width (l/w) 2
2 Bag pressure to cushion pressure (Pb/Pc) 1.3
Sensors 2018, 18, 3318 5 of 23

Table 1. Assumptions of the hovercraft parameters.

Serial No. Empirical Relation Limit

1 Length to width (l/w) 2
2 Bag pressure to cushion pressure (Pb /Pc ) 1.3
3 Forward thrust to overall weight during hovering (Tf /W) 0.2
4 Propeller pitch to diameter (p/d) 0.6
5 Vertical thrust to maximum take-off weight (Tv /W) 2

Table 2. Calculation of the hovercraft parameters.

Parameter Empirical Relation Values

Maximum take-off weight (W) m×g 269.78 N
Length of the hovercraft (l) 2×w 1.00 m
Cushion Area (Ac ) l × w − πr2 0.40 m2
W
Cushion pressure (Pc ) 674.44 N/m2
qAc
Air escaping velocity (Ve ) 2 Pρc 33.18 m/s
Air escaping area (Ae ) 2 × (l + w) × h 0.038 m2
Air flow rate (Qe ) Ae × Ve 1.26 m3 /s
Power required (Pe ) Qe ×ρ×Ve 2 852.32 W
2
m—mass of the hovercraft; w—width of the hovercraft; l—length of the hovercraft; ρ —air density;
h—hovering height.

3.1.1. Selection of Skirt Material

The inflation of skirt functionality for realization of hovering means the skirt material should have
sufficient tensile strength. A survey of various skirt materials [21–25] for hovercrafts given in Table 3
reveals that nylon impregnated with urethane was the best choice due to its high tensile strength, light
weight, and superior resistance to wear and tear characteristics.

Table 3. Mechanical properties of skirt.

Nylon Impregnated Natural Vinyl-Coated 1000

Property
with Urethane Rubber-Coated Nylon Denier Polyester
Tensile strength (MPa) 45 35 3.06
Elastic modulus (MPa) 1.48 1.10 20.00
Density (kg/m3 ) 900 1016 1500
Hardness 75 34 87
Flexural strength (MPa) 41 20 26

3.1.2. Selection of Hull Material

The hull is considered to be a watertight body part of the hovercraft and it has to support various
payloads, the battery, and other electronic systems. It has to withstand high upward pressure generated
through the cushion of air during the inflation of the skirt. Based on the survey of material [21–25]
options (Table 4), polyurethane foam was selected due to its high strength-to-weight ratio.

Table 4. Mechanical properties of hull.

Property Composite Material Fiberglass Polyurethane Foam

Tensile strength (MPa) 1200 1950 1900
Compressive strength (MPa) 866 4000 48
Elastic modulus (GPa) 45 72 4
Density (kg/m3 ) 7850 2540 1390
Flexural strength (MPa) 146 110 57
Sensors 2018, 18, 3318 6 of 23

3.2. Multicopter Design

A hollow square cross-section aluminium channel was considered for the horizontal and vertical
frames for supporting the pair of motors. The speed and thrust of the motors [26] were calculated
(Table 5) using empirical relations.

Sensors 2018, 18, 3318 Table 5. Multi-rotor structure parameters. 6 of 22

Parameters Table 5. Multi-rotorEmpirical

structure parameters.
Relation Values
q
Parameters
Required motor speed (N) ×1010
Lm
Empirical Relation Values
4200 rpm
p×d3 ×0.0283495× g
Thrust per motor p × d3 × N2 −
× 10 𝐿×10 𝑚× 1010 ×
Required motor(T)
speed (N) √
0.0283495 g 92 Nrpm
4200
Lift required for the multicopter (Lm ) 𝑝×
2×𝑑3W
× 0.0283495 × 𝑔 540 N
Thrustlift
Lm —multicopter perrequired;
motor (T) × 𝑑3 × 𝑁 2 diameter;
p—propeller pitch;𝑝d—propeller × 10−10 × 0.0283495 × 𝑔due to gravity.
g—acceleration 92 N
Lift required for the multicopter (Lm) 2 ×W 540 N
Lm—multicopter lift required; p—propeller pitch; d—propeller diameter; g—acceleration due to gravity.
3.2.1. Propulsion
3.2.1. Propulsion
As perAsthe initial estimation of the speed of the motor, the motor was selected with reference to
per the initial estimation of the speed of the motor, the motor was selected with reference to
the Kvthe
rating
Kv rating Kv)
(125 (125 having a power
Kv) having a powerof of
1900
1900W.
W.Table
Table66illustrates thenecessary
illustrates the necessary current
current rating
rating and and
number of cells
number of of a battery
cells for 25
of a battery for min of of
25 min endurance
endurancefor
foraabattery capacityofof22,000
battery capacity 22,000 mAh.
mAh.

Table 6. Propulsion
Table 6. Propulsionsystem
systemparameters.
parameters.

Parameters
Parameters Empirical Relation
Empirical Relation Values Values
Speed of motor (N) 𝐾𝑣 × 𝑉 4800 rpm
Speed of motor (N) Kv × V 𝑃 4800 rpm
Operating current (I) P 50 A
Operating current (I) V 𝑉 50 A
Number ofNumber
batteryofcells (nS)cells (nS) V 𝑉 10 10
battery 3.7
Endurance (E) mAh × 0.001 × 60/Σ Im3.7 25 min
Endurance (E) mAh × 0.001 × 60/Σ Im 25 min
Kv—motor
Kv—motor V—operating
rating;rating; voltage;
V—operating P—motor
voltage; power;mAh—milliampere
power;
P—motor hour; IImm—summation
mAh—milliamperehour; —summation of
current consumption.
of current consumption.

3.2.2. Selection of Propeller

3.2.2. Selection of Propeller

The weight of amphibious

The weight of amphibious system
systemwaswasconsidered
considered totobebe3030
kgkg
forfor selection
selection of theofpropellers.
the propellers.
A
A Quad Quad with
with a co-axial
a co-axial motor–propeller configuration
motor–propeller configuration was
wasconsidered
considered due
dueto to
thethe
demand
demand of high
of high
payload-carrying
payload-carrying capacity
capacity and stability
and stability of vehicle.
of the the vehicle. Considering
Considering thethe thrusttotoweight
thrust weightratio
ratioas
as2%
2% and
and a 20% thrust loss due to the co-axial configuration, the maximum thrust was estimated as 75 kg.
a 20% thrust loss due to the co-axial configuration, the maximum thrust was estimated as 75 kg. Under
Under the full-throttle condition, propellers of various diameters and their thrust force [27]
the full-throttle condition, propellers of various diameters and their thrust force [27] characteristics
characteristics were examined (Figure 3). In order to lift 75 kg, each co-axial arm needed to produce
were examined (Figure 3). In order to lift 75 kg, each co-axial arm needed to produce approximately
approximately 18.5 kg of thrust. Hence, a co-axial propeller configuration with diameter 0.75–0.80 m
18.5 kgwas
of thrust.
selected.Hence, a co-axial propeller configuration with diameter 0.75–0.80 m was selected.

60

50

40
Thrust (kg)

30

20

10

0
0.4 0.6 0.8 1 1.2
Propeller diameter (m)

Figure 3. Selection
Figure ofofthe
3. Selection theco-axial
co-axialpropeller diameter.
propeller diameter.

3.2.3. Selection of the Motor

Selection of the motor primarily depends upon the size of the propeller for generating sufficient
thrust, and the selected configuration demands 10 kg thrust force per motor. Power consumption of
Sensors 2018, 18, 3318 7 of 23

3.2.3. Selection of the Motor

Selection
Sensors of the motor primarily depends upon the size of the propeller for generating sufficient
2018, 18, 3318 7 of 22
thrust, and the selected configuration demands 10 kg thrust force per motor. Power consumption of
the 125
125 kvkv motor
motor under
under various
various speeds
speeds was
was determined
determined as
as shown
shown in
in Figure
Figure 4.
4. Under
Under aa full-throttle
full-throttle
condition, aa maximum
condition, maximum power
power of
of 1.86
1.86 kW
kW is
is required
required per
per each
each motor
motorat
ataaspeed
speedofof4480
4480rpm.
rpm.

2
125kv motor
1.8
1.6
1.4
Power (kW)

1.2
1
0.8
0.6
0.4
0.2
0
0 1000 2000 3000 4000 5000
Motor speed (rpm)

Figure 4. Power
Figure 4. Power consumption
consumption characteristics
characteristics of
of the
the selected
selected motor.
motor.

In
In this
this work,
work, aa T-Motor
T-Motor MN705-S
MN705-S (T-Motor,
(T-Motor, Nanchang,
Nanchang, China) with 125
China) with 125 Kv,
Kv, ideal
ideal current of 1.4
current of 1.4 A,
A,
peak current of 45 A, and maximum power of 2200 W was considered [28].
peak current of 45 A, and maximum power of 2200 W was considered [28].
3.2.4. Selection of the Battery
3.2.4. Selection of the Battery
Selection of a battery depends upon the consumption of current with sufficient voltage and
Selection of a battery depends upon the consumption of current with sufficient voltage and
discharge rate requirements. Total current consumption for the electronic components was calculated
discharge rate requirements. Total current consumption for the electronic components was calculated
to be 2.73 A (Table 7). Considering eight motors, the total power and current required for the vehicle
to be 2.73 A (Table 7). Considering eight motors, the total power and current required for the vehicle
to fly in the air were estimated as 3.78 kW and 79.13 A (Table 8), respectively. However, when the
to fly in the air were estimated as 3.78 kW and 79.13 A (Table 8), respectively. However, when the
vehicle lands on water and glides along the water surface, two pairs of motors need to be actuated.
vehicle lands on water and glides along the water surface, two pairs of motors need to be actuated.
Estimation of current and power consumption during the gliding of vehicle on the water surface is
Estimation of current and power consumption during the gliding of vehicle on the water surface is
given in Table 9 and it is evident that only half of the power is required for the amphibious mode as
given in Table 9 and it is evident that only half of the power is required for the amphibious mode as
compared to the flight mode.
compared to the flight mode.
Table 7. Estimation of power and current of on-board electronics.
Table 7. Estimation of power and current of on-board electronics.
Power Required Current Consumption
Sl. No Component Name Power Required Current Consumption
Sl. No Component Name (W) (A)
(W) (A)
1 1 Flight controller
Flight controller board board 16 16 0.34 0.34
2 2 ESC ESC 40 40 0.83 0.83
3 Video telemetry 15 0.34
3 Video telemetry 15 0.34
4 Camera 10 0.21
4 Camera 10 0.21
5 On-board processor 50 1.01
5 On-board processor 50 1.01
Total
Total 131 131 2.73 2.73

Table 8. Power consumption (airborne mode).

Total Power Total Current

Power
Sl. No Components Consumption Consumption
(W)
(W) (A)
1 Electronics components 95.30 95.30 2.73
3686.40 (for eight 76.40 (for eight
2 Motors 460.80 (per motor)
motors) motors)
Total 3781.70 79.13
Sensors 2018, 18, 3318 8 of 23

Table 8. Power consumption (airborne mode).

Total Power Total Current

Sl. No Components Power (W)
Consumption (W) Consumption (A)
1 Electronics components 95.30 95.30 2.73
2 Motors 460.80 (per motor) 3686.40 (for eight motors) 76.40 (for eight motors)
Total 3781.70 79.13

Table 9. Power consumption (gliding on water).

Total Power Total Current

Sl. No Components Power (W)
Consumption (W) Consumption (A)
1 Electronics components 95.30 95.30 2.73
2 Motors 240.00 (per motor) 1920.00 (for four motors) 40.00 (for four motors)
3 Ducted Motor 100.00 (per motor) 100.00 (for one motor) 2.08 (for one motor)
Total 2115.30 44.81

In order to meet these power and current consumption requirements, a 22,000 mAh capacity
battery was selected. It was expected to have an estimated endurance of 22 min while in air flight and
46 min during hovering or gliding along the water body.

3.3. Weight Estimation

Based upon earlier selection of materials for hovercraft and multirotor components, weight of the
amphibious vehicle is estimated as 27.31 kg inclusive of 7 kg payload. Table 10 shows the weight of
various components of hovercraft and multirotor systems.

Table 10. Weight of each components

Sl. No. Components Weight (kg)

1 Multicopter frames (aluminum alloy 6061) 1.40
2 Hull (Polyurethane foam) 0.80
3 Skirt (Nylon impregnated with urethane) 1.70
4 Control system 0.45
5 Multicopter motor 3.36
6 Multicopter propeller 0.31
7 Multicopter electronic speed controller (ESC) 0.88
8 Servo 0.50
9 Electronic duct fan (EDF) 0.40
10 EDF ESC 0.11
11 Li-Po batteries 10.00
12 Miscellaneous 0.40
13 Payload 7.00
Total Weight 27.31

4. Structural Analysis of the Amphibious Vehicle

The multirotor configuration had vertical and horizontal frames that were made of aluminium
channels owing to its lightweight characteristics. At the tip of the horizontal frames, the motor and
co-axial propellers were attached. The vertical frames were anchored to the top surface of the hovercraft
hull. The thrust produced by the propellers [29–31] was considered to be acting at the fixed support of
horizontal frame and the same vertical axial loading was applied at the four corners of the horizontal
frame. An axial load was applied at the tip of the frame and the effect of cushion pressure generated
through the duct fan located at the center of hull was analyzed. The pressure load was applied at the
inner surface of the skirt and bottom of the hull. The effect of these loading conditions was evaluated
through structural analysis. A displacement of 0.6 mm was experienced at the tip of horizontal frame
Sensors 2018, 18, 3318 9 of 23

(Figure 5). The von Mises stress plot (Figure 6) shows that the junction of horizontal and vertical
frames experiences maximum stress regions about 25 MPa. Other portions of the amphibious structure
experienced
Sensors 2018, 18,considerably
3318 lower levels of stress. 9 of 22
Sensors 2018, 18, 3318 9 of 22

Figure
Figure 5.
Figure 5. Deformation
5. Deformation plot
Deformation plot of
plot of the
of the amphibious
the amphibious structure.
amphibious structure.

Figure
Figure 6.
Figure 6. Stress
6. Stress contour
Stress contour of
contour of the
of the amphibious
the amphibious structure.
amphibious structure.
structure.
5. Aerodynamic Analysis
5. Analysis
5. Aerodynamic
Aerodynamic Analysis
Aerodynamic evaluation of
Aerodynamic of the amphibious
amphibious vehicle was was performed throughthrough varying
varying thethe wind
Aerodynamic evaluation
evaluation of the the amphibious vehicle
vehicle was performed
performed through◦ and 8◦varying the wind
wind
speeds
speeds in the range of 5 to 10 m/s with different angles of attack (AOA) (0 ). Computational
speeds inin the
the range
range of
of 55 to
to 10
10 m/s
m/s with
with different
different angles
angles ofof attack
attack (AOA)
(AOA) (0° (0° and
and 8°).
8°). Computational
Computational
fluid dynamic (CFD)
fluid analysis using the ANSYS
ANSYS FLUENT platformplatform (ANSYS, Canonsburg,
Canonsburg, PA, USA) USA) was
fluid dynamic
dynamic (CFD)
(CFD) analysis
analysis using
using the
the ANSYS FLUENT
FLUENT platform (ANSYS,
(ANSYS, Canonsburg, PA, PA, USA) was
was
used
used to examine the velocity and pressure contours during forward flight conditions, and aerodynamic
used to
to examine
examine thethe velocity
velocity and
and pressure
pressure contours
contours during
during forward
forward flight
flight conditions,
conditions, and
and aerodynamic
aerodynamic
coefficients were also
coefficients also determined. The The quality ofof meshing (Figure
(Figure 7) isis evaluated through
through performing
coefficients were
were also determined.
determined. The quality
quality of meshing
meshing (Figure 7) 7) is evaluated
evaluated through performing
performing
orthogonality
orthogonality and skewness characteristics (0.9). Inlet as velocity and outlet as awaspressure was
orthogonality and skewness characteristics (0.9). Inlet as velocity and outlet as a pressure was considered
and skewness characteristics (0.9). Inlet as velocity and outlet as a pressure considered
considered
and and boundaries were defined far away (10 times) to reduce the horizontal buoyancy effect
and boundaries
boundaries werewere defined
defined farfar away
away (10
(10 times)
times) to
to reduce
reduce the
the horizontal
horizontal buoyancy
buoyancy effect
effect and
and wall
wall
and wall
inference. inference. A symmetric plane and no heat transfer were assumed to perform simulations.
inference. A A symmetric
symmetric plane
plane and
and no
no heat
heat transfer
transfer were
were assumed
assumed to to perform
perform simulations.
simulations.
Sensors 2018, 18, 3318 10 of 23

Sensors 2018, 18, 3318 10 of 22

Sensors 2018, 18, 3318 10 of 22

Figure
Figure 7.
Figure 7. Meshed
7. Meshed geometry.
Meshed geometry.
geometry.

Simulation
Simulation results
results indicated
indicated that
that at
at various
various angles
angles of
of attack,
attack, collision
collision of
collision of air
air with
air with the
with the frontal
the frontal body
frontal body
body
surface
surface caused a velocity drop (Figure 8) due to stagnation pressure (Figure 9), and there was a loss
caused a velocity
velocity drop
drop (Figure
(Figure 8)
8) due
due toto stagnation
stagnation pressure
pressure (Figure
(Figure 9),9),
andand there
there waswasa a
loss of
loss
of kinetic
kinetic energy.
energy.
of kinetic At
AtAt
energy. the
thethe middle
middle
middle of
ofof the
the amphibious
theamphibious vehicle,
amphibiousvehicle, a low-pressure
vehicle,aalow-pressure region
low-pressureregion formed
regionformed that created
formed that created
vortex
vortex and flow separation. This phenomenon may have created an imbalance of
and flow separation. This
This phenomenon
phenomenon may
may have
have created
created an imbalance
imbalance of the
the vehicle,
vehicle, which
which
can
can be streamlined through providing riblets. At the rear of the vehicle, recirculation flow
can be streamlined
streamlined through
through providing
providing riblets.
riblets. At
At the
the rear
rear of
of the vehicle, recirculation flow occurred
occurred
due
due to
to non-uniformity
non-uniformity and and aa blunt
blunt profile
profile of
of the UAV
the UAV structure.
UAVstructure.
structure.

Figure 8. Velocity
Velocity streamline at
at 44°
streamline at ◦ AoA.
4°
Figure 8. Velocity streamline AoA.

Figure
Figure 9.
9. Pressure
Pressure contour
contour at
at 4°
4° AoA.
AoA.
Sensors 2018, 18, 3318 11 of 23
Figure 8. Velocity streamline at 4° AoA.

Figure
Figure 9.
9. Pressure
Pressure contour
contour at 4◦ AoA.
at 4° AoA.
Sensors 2018, 18, 3318 11 of 22
For various angles of attack, the co-efficient of drag was estimated and the corresponding drag
For various angles of attack, the co-efficient of drag was estimated and the corresponding drag
force was calculated (Table 11). It is evident that substantial drag was experienced, which reduced the
force was calculated (Table 11). It is evident that substantial drag was experienced, which reduced
endurance of the UAV. In order to reduce the effect of drag, an inclined front panel (Figure 10) and
the endurance of the UAV. In order to reduce the effect of drag, an inclined front panel (Figure 10)
blended nose configurations (Figure 11) were considered.
and blended nose configurations (Figure 11) were considered.
Table 11. Drag Estimation.
Table 11. Drag Estimation.
α Cd D
α Cd D
Angle of Attack Drag Coefficient Drag (N)
Angle of Attack Drag Coefficient Drag (N)
0◦ 0° 5.89 5.89 38.3 38.3
4◦ 5.65 36.8
4° 5.65 36.8
8◦ 5.50 35.8
8° 5.50 35.8

Figure 10.
Figure 10. Flat
Flat panel—Pressure
panel—Pressure contour.
contour.

Figure 11. Blended nose—Pressure contour.

Sensors 2018, 18, 3318 12 of 23
Figure 10. Flat panel—Pressure contour.

Figure
Figure 11.
11. Blended
Blended nose—Pressure contour.
nose—Pressure contour.

6. Water Sample
6. Water Sample Collection
Collection Using
Using Robotic
Robotic Arm
Arm
A two-degree-of-freedom (DOF) manipulator actuated using a servo motor was used to collect
water
water samples
samples as as shown
shown in in Figure
Figure 12.
12. An
An end-effector
end-effector carried
carried aa water-sucking
water-sucking pump,pump, which
which inin turn
turn
connected
connected through
througha ahose.
hose.Drawn
Drawn water waswas
water collected in theinrespective
collected storagestorage
the respective tank with a 1with
tank L capacity.
a 1 L
The depthThe
capacity. of water
depth collection
of water was controlled
collection wasusing a rope driven
controlled usingby a arope
stepper motor.
driven by Encoder
a stepperfeedback
motor.
was sent feedback
Encoder to an Arduino
was sentbased controller
to an Arduino(Arduino Srl, Strambino,
based controller (Arduino Italy) to monitor Italy)
Srl, Strambino, the depth of the
to monitor
collection
the depth of the collection of water. The arm of robot manipulator was made up of carbon fiberservo
of water. The arm of robot manipulator was made up of carbon fiber and a waterproof and
motor was attached
a waterproof servo at each was
motor link of the robotic
attached arm.link
at each During
of thetherobotic
water sample collection,
arm. During the stability of the
water sample
vehicleSensors
collection, 2018, 18, 3318
wasstability
assured through
of distributing
the vehicle waterthrough
was assured using a two-way
distributingcontrol
water value.
usingA awater sensor12 of 22
levelcontrol
two-way
was used
value. to measure
A water level the quantity
sensor of water, and corresponding feedback was sent to control the pump
was sent to control thewas
pumpused to retraction
and measure the quantity
of the robotic ofarm.
water, and plates
Buffer corresponding
were placedfeedback
in the water
and retraction of the robotic arm. Buffer plates were placed in the water storage tank to dampen the
storage tank to dampen the vibration caused due to turbulence of water in the storage tank.
vibration caused due to turbulence of water in the storage tank.

(a) (b)

Figure 12. Robotic arm with suction pump assembly, (a) Folded arm (b) Extended arm.
Figure 12. Robotic arm with suction pump assembly, (a) Folded arm (b) Extended arm
Figure 13 illustrates the payload control unit, in which pulse-width modulated (PWM) signals
Figure
were sent to 13 the
actuate illustrates the payload
servo motors. The control unit,sensors,
water level in whichwater
pulse-width
quality modulated
monitoring(PWM)
sensors,signals
pump,were sent to actuate
and encoder were usedtheto
servo motors.
provide The water
feedback level
in analog andsensors,
digitalwater
forms.quality monitoring sensors,
pump, and encoder were used to provide feedback in analog and digital forms.
 Figure 12. Robotic arm with suction pump assembly, (a) Folded arm (b) Extended arm

Figure 13 illustrates the payload control unit, in which pulse-width modulated (PWM) signals
were sent to actuate the servo motors. The water level sensors, water quality monitoring sensors,
pump,2018,
Sensors and18,encoder
3318 were used to provide feedback in analog and digital forms. 13 of 23

Figure 13. Main

Figure 13. Main control
control unit.
unit.
7. Development of Ground Control Station
The ground control station (Figure 14) consisted of a portable computer with payload
control, mission management, and flight data monitoring with corresponding communication links.
The mission of the amphibious vehicle was pre-planned in the ground control station wherein the
vehicle was flown as multi-rotor to identify contaminated regions of the waterbody and on-line video
was streamed using a 5.8 GHz video data link. Once contaminated regions are identified, the vehicle
was landed on the water surface through the hovercraft mode. Water samples were collected using a
robotic arm with suction pump and rope mechanism structure. Radio frequency signals in universal
asynchronous receiver—transmitter (UART) carrier mode were used to communicate and actuate the
servos, sensors, and other actuators to collect required water samples with precise feedback.
was flown as multi-rotor to identify contaminated regions of the waterbody and on-line video was
streamed using a 5.8 GHz video data link. Once contaminated regions are identified, the vehicle was
landed on the water surface through the hovercraft mode. Water samples were collected using a
robotic arm with suction pump and rope mechanism structure. Radio frequency signals in universal
asynchronous receiver—transmitter (UART) carrier mode were used to communicate and actuate the
Sensors 2018, 18, 3318 14 of 23
servos, sensors, and other actuators to collect required water samples with precise feedback.

Figure 14. Ground

Figure 14. Ground control
control station.
station.

A typical flight control computer [32] is presented in Figure 15. It acted as the central hub of
A typical flight control computer [32] is presented in Figure 15. It acted as the central hub of the
the system through which position, orientation, and heading direction of the vehicle were controlled.
system through which position, orientation, and heading direction of the vehicle were controlled. In
In addition, the receiving and transmission of data, battery power monitoring, and actuation of servos,
addition, the receiving and transmission of data, battery power monitoring, and actuation of servos,
motors, and the robotic arm were performed.
motors, and the robotic arm were performed.
The airborne mode of the mission is depicted in Figure 16 and explains that the radio frequency
The airborne mode of the mission is depicted in Figure 16 and explains that the radio frequency
signals at 2.4 GHz were transmitted and received through telemetry modules. The received signals are
signals at 2.4 GHz were transmitted and received through telemetry modules. The received signals
sent in pulse position modulated (PPM) form to the flight controller. The flight controller computer
are sent in pulse position modulated (PPM) form to the flight controller. The flight controller
handled control and navigation of the vehicle during flying and hovering modes. PWM signal from
computer handled control and navigation of the vehicle during flying and hovering modes. PWM
flight control computer is sent to the electronic speed controller (ESC) to actuate the brushless direct
signal from flight control computer is sent to the electronic speed controller (ESC) to actuate the
current (BLDC) motor to lift and navigate the vehicle in the desired attitude and altitude. The on-line
brushless direct current (BLDC) motor to lift and navigate the vehicle in the desired attitude and
video streaming was achieved through RF mode and on-screen display module is integrated to monitor
altitude. The on-line video streaming was achieved through RF mode and on-screen display module
the water surface in real time during flying mode. Autonomous capability was achieved through way
is integrated to monitor the water surface in real time during flying mode. Autonomous capability
point navigation, guidance, and control with prior mission planning.
was achieved through way point navigation, guidance, and control with prior mission planning.
Sensors 2018, 18, 3318 14
15of
of 22
23
Sensors 2018, 18, 3318 14 of 22

Figure
Figure 15. Flight
Flight control
control computer
computer module.
module.
Figure 15. Flight control computer module.

Figure 16. Airborne mission.

Figure 16. Airborne mission.
Figure 16. Airborne mission.
During hover mode of the vehicle (Figure 17), the payload control unit was triggered to actuate
Duringarm,
the robotic hover mode
water of theand
pump, vehicle
water(Figure
quality17), the payload
sensors control
to collect waterunit was triggered
samples to actuate
and perform in situ
the robotic arm, water
water quality analysis. pump, and water quality sensors to collect water samples and perform in situ
water quality analysis.
Sensors 2018, 18, 3318 16 of 23

During hover mode of the vehicle (Figure 17), the payload control unit was triggered to actuate
the robotic arm, water pump, and water quality sensors to collect water samples and perform in situ
water quality
Sensors 2018, analysis.
18, 3318 15 of 22

Figure 17.
Figure Water sampling
17. Water sampling mission.
mission.
8. Fabrication and Assembly of Amphibious Structure
8. Fabrication and Assembly of Amphibious Structure
An amphibious vehicle structure was fabricated (Figure 18) based upon the selected motors,
An amphibious vehicle structure was fabricated (Figure 18) based upon the selected motors,
propellers, battery, hull, and skirt materials. On top of the hull, vertical hollow aluminum frames were
propellers, battery, hull, and skirt materials. On top of the hull, vertical hollow aluminum frames
mounted upon which horizontal frames were fixed. At the four corners of the horizontal frame, 3-D
were mounted upon which horizontal frames were fixed. At the four corners of the horizontal frame,
printed knuckle joints were used and a motor-propeller configuration was mounted on it. A servo
3-D printed knuckle joints were used and a motor-propeller configuration was mounted on it. A servo
motor is attached to rotate the motor-propeller configuration. At the center of the hull, a propeller was
motor is attached to rotate the motor-propeller configuration. At the center of the hull, a propeller
mounted that produced the necessary pressure to lift the vehicle through inflating the skirt. An open
was mounted that produced the necessary pressure to lift the vehicle through inflating the skirt. An
source advanced level controller was utilized to control and navigate the vehicle. The constructed
open source advanced level controller was utilized to control and navigate the vehicle. The
amphibious vehicle was tested in an ambient environment and stable flight was observed (Figure 19a).
constructed amphibious vehicle was tested in an ambient environment and stable flight was observed
Preliminary testing of the vehicle was also taken up in a water tank (Figure 19b) and water was
(Figure 19a). Preliminary testing of the vehicle was also taken up in a water tank (Figure 19b) and
collected through actuating the suction pump.
water was collected through actuating the suction pump.
 motor is attached to rotate the motor-propeller configuration. At the center of the hull, a propeller
was mounted that produced the necessary pressure to lift the vehicle through inflating the skirt. An
open source advanced level controller was utilized to control and navigate the vehicle. The
constructed amphibious vehicle was tested in an ambient environment and stable flight was observed
(Figure 19a). Preliminary testing of the vehicle was also taken up in a water tank (Figure 19b) and
Sensors 2018, 18, 3318 17 of 23
water was collected through actuating the suction pump.

Sensors 2018, 18, 3318 Figure 18. Prototype

Figure 18. Prototype of
of amphibious
amphibious UAV.
UAV. 16 of 22

(a) (b)

Figure
Figure 19.
19. Field
Fieldtesting
testingof
ofamphibious
amphibiousUAV:
UAV: (a)
(a) Amphibious
AmphibiousUAV
UAV (airborne);
(airborne); (b)
(b) Amphibious
AmphibiousUAV
UAV
(gliding
(gliding above
above water).
water).

The folded
The folded robotic
robotic arm
arm was
was extended
extended through
through aa servo
servo actuator
actuator after
after landing
landing the
the vehicle
vehicle on
on the
the
water surface
water surface and
and aa rope
rope mechanism
mechanism attached
attached with
with aa suction
suction pump
pump was
was actuated
actuated toto suck
suck the
the water
water
at the
at the desired
desired depth
depth ofof the
the water
water channel.
channel. After
After performing
performing thethe waterborne
waterborne mission,
mission, the
the robotic
robotic arm
arm
assembly was
assembly wasretracted
retracted totothe
theinitial
initialfolded
foldedconfiguration
configurationfor
forcompactness
compactnessandandstability
stabilityofofthe
thevehicle.
vehicle.

9. Internet
9. Internet of
of Things
Things (IoT)-Based
(IoT)-Based Water
Water Quality
Quality Measurement
Measurement

ItIt is
isessential
essentialto toperform
performwaterwaterquality
qualityinspection
inspection atatregular
regular intervals
intervals at
atthe
thewater
waterreservoirs
reservoirs such
such
as dams,
as dams, lakes,
lakes, rivers,
rivers, and
and ponds.
ponds. Collection
Collectionofof water
water samples
samples in in remote
remote water
water bodies
bodies is is challenging
challenging
and time consuming. Traditional methods of collecting water samples
and time consuming. Traditional methods of collecting water samples using boats is cumbersomeusing boats is cumbersome
and itit is
and is very
very difficult
difficult to
to access
access remote
remote water
water locations.
locations. InIn this
this work,
work, wewe developed
developed an an amphibian
amphibian
vehicle that
vehicle thatcancanmeasure
measurethe thewater
water quality
quality using
using various
various on-board
on-board water
water quality
quality sensors
sensors suchsuch as
as pH,
pH, dissolved oxygen (DO), electrical conductivity (EC), temperature, and
dissolved oxygen (DO), electrical conductivity (EC), temperature, and turbidity, and they were turbidity, and they were
procured. A
procured. ARaspberry
Raspberrypipi zero BCM
zero BCM2835,2835,
1 GHz 1 ARM
GHz 11 ARMcore,11512 MB of
core, 512LPDDR2
MB of SDRAM
LPDDR2(Raspberry
SDRAM
PI Foundation, Cambridge. UK) was utilized to process the sensor
(Raspberry PI Foundation, Cambridge. UK) was utilized to process the sensor data and data and sent via a 4Gsent dongle-LTE
via a 4G
network at 2300
dongle-LTE MHz. at
network The IoT setup
2300 MHz. shown
The IoTin Figure
setup 20 was embedded
shown in Figure into the designed
20 was embedded amphibious
into the
UAV andamphibious
designed in situ waterUAV quality
and inmeasurements weremeasurements
situ water quality performed. The weresensor data were
performed. sampled
The sensor at
data
a 160 MHz sampling speed using an Arduino pro mini and transmitted
were sampled at a 160 MHz sampling speed using an Arduino pro mini and transmitted to the to the Raspberry pi UART
section. The
Raspberry pi Raspberry pi was
UART section. Theconnected
Raspberry with a 4G-LTE
pi was dongle,
connected withand the UAVdongle,
a 4G-LTE was operated
and thewith
UAVa
2.4 GHz radio frequency to avoid inference.
was operated with a 2.4 GHz radio frequency to avoid inference.

Raspberry pi
with Wi-Fi
router
 procured. A Raspberry pi zero BCM 2835, 1 GHz ARM 11 core, 512 MB of LPDDR2 SDRAM
(Raspberry PI Foundation, Cambridge. UK) was utilized to process the sensor data and sent via a 4G
dongle-LTE network at 2300 MHz. The IoT setup shown in Figure 20 was embedded into the
designed amphibious UAV and in situ water quality measurements were performed. The sensor data
were sampled at a 160 MHz sampling speed using an Arduino pro mini and transmitted to the
Raspberry
Sensors 2018, 18, 3318 pi UART section. The Raspberry pi was connected with a 4G-LTE dongle, and the UAV18 of 23
was operated with a 2.4 GHz radio frequency to avoid inference.

Raspberry pi
with Wi-Fi
router

Water quality
sensors
(pH, DO, EC,
turbidity)

Figure
Figure 20. Sensor
20. Sensor interface
interface with
with Raspberrypi.
Raspberry pi.

The The18,sensor
Sensorssensor
2018, module
3318 module encompassed
encompassed pH,pH, dissolvedoxygen
dissolved oxygen(DO),
(DO), electrical
electrical conductivity
17 of (EC),
conductivity 22(EC),
temperature, and turbidity, which was immersed into the collected water sample during the
temperature, and turbidity, which was immersed into the collected water sample during the
deployment of the amphibious vehicle in a desired location. Initially, water quality sensors were
deployment
calibrated ofwith
the corresponding
amphibious vehicle in a desired
acidic solutions. The pHlocation. Initially,
sensor was immersedwater
in aquality sensors
KCl solution, thewere
calibrated with corresponding acidic solutions. The pH sensor was immersed
DO sensor tip was stored in NaoH solution, and EC was calibrated with distilled water. These in a KCl solution,
the DO sensor tip
calibrations was
were stored infor
performed NaoH
each solution,
mission beforeand EC wasmeasurement.
in situ calibrated withOncedistilled
the waterwater.
quality These
calibrations were performed
was measured for sample,
for a particular each mission
water wasbefore in situ
drained andmeasurement.
a new water sampleOnce thecollected
was water quality
in
a container.for
was measured The sensors were
a particular flushedwater
sample, with distilled water for
was drained andeach measurement
a new water sampleto obtain
wasaccurate
collected in
results. The measured analog signal was sent to the embedded computer
a container. The sensors were flushed with distilled water for each measurement to obtainas 16 bit digital data. accurate
A typical
results. The measuredIoT-based
analognetwork
signal is shown
was sentintoFigure 21, which demonstrates
the embedded computer as the working
16 bit principle
digital data.
of the UAV-based water quality measurement and transmission of data.
A typical IoT-based network is shown in Figure 21, which demonstrates the working principle of
the UAV-based water quality measurement and transmission of data.

Figure
Figure 21.21.
IoTIoTarchitecture
architecture for
for water
water quality
qualitymeasurement.
measurement.

The ARM-based computer processed the sensor data and provided the useful information for
measurable quantities in standard units. The water quality information was transferred to the cloud
database for real-time monitoring and post-processing. The link between the cloud data base server
and real-time on-board embedded computer were created with the 4G broadband cellular network
because it provided better speed in comparison to 2G and 3G. The water quality data in the cloud
Sensors 2018, 18, 3318 19 of 23

The ARM-based computer processed the sensor data and provided the useful information for
measurable quantities in standard units. The water quality information was transferred to the cloud
database for real-time monitoring and post-processing. The link between the cloud data base server and
real-time on-board embedded computer were created with the 4G broadband cellular network because
it provided better speed in comparison to 2G and 3G. The water quality data in the cloud could be
accessed through the smart devices with an Internet service anywhere in the world. Preliminary
experiments were conducted to examine the performance characteristics of the sensor network.
The saturation time taken for each sensor was obtained and turbidity took more time to arrive
Sensors 2018, 18, 3318 18 of 22
at saturation
Sensors 2018, 18,in comparison with other sensors, as shown in Figure 22.
3318 18 of 22
70 65
70 65
(ms)

60
(ms)

60
50
time

50
time

40
31
Saturation

40 30
26 31
Saturation

30 30
30 26
20
20
10
10
0
0 pH Dissolved Turbidity Electrical
pH Dissolved
oxygen Turbidity conductivity
Electrical
oxygen
Water quality sensorsconductivity
Water quality sensors
Figure 22. Sensor saturation time.
Figure 22. Sensor saturation time.
Figure 22. Sensor saturation time.
The average
The average delay
delay (500
(500 set
set transmissions)
transmissions) in
in water
water quality
quality data
data was
was monitored
monitored with
with different
different
The average
network delay
conditions (4G, (500
3G, set
andtransmissions)
2G). The in waterdelay
round-trip quality data
time was monitored
seemed to be with
better fordifferent
4G-LTE
network conditions (4G, 3G, and 2G). The round-trip delay time seemed to be better for 4G-LTE
network conditions
communication, and(4G,
it 3G,11and
took s on2G). The to
average round-trip
reach thedelay time seemed
destination (Figure to be better for 4G-LTE
23).
communication, and it took 11 s on average to reach the destination (Figure 23).
communication, and it took 11 s on average to reach the destination (Figure 23).

20 17
trip

20 17
trip

15
(S)(S)

11 12
round

15 12
time

11
round

10
time

10
Average

5
delay
Average

5
delay

0
0 4G-LTE 3G 2G
4G-LTECommunication
3G Network
2G
Communication Network
Figure 23. Round trip time.
Figure 23. Round trip time.
Figure 23. Round trip time.
In addition, power consumption of various sensors, and the transmitting and receiving unit was
In addition, power consumption of various sensors, and the transmitting and receiving unit was
calculated and it is shown in Figure 24.
In addition,
calculated and it power consumption
is shown of various sensors, and the transmitting and receiving unit was
in Figure 24.
The transmitted sensor data was collected in the Google firebase cloud as shown in Figure 25 and
calculated and it is shown in Figure 24.
The transmitted sensor data was collected in the Google firebase cloud as shown in Figure 25
it could be synchronized with the cloud messaging service and shared across the globe through the
and itThe transmitted
could sensor data
be synchronized withwas
the collected in the Google
cloud messaging servicefirebase cloud
and shared as shown
across in Figure
the globe 25
through
firebase cloud services.
and it could be synchronized
the firebase cloud services. with the cloud messaging service and shared across the globe through
the firebase cloud services.

Power consumption(W) High Power

Power consumption(W) High
Low Power
3 Power consumption(W) Low Power
2.7
(W)

3 2.48 2.48 2.7

(W)

2.5 2.48 2.48

Consumption

2.5
ronsumption

2
2 1.5
1.5 1.5
1.5 1.1
1 1.1
1 0.5
 In addition, power consumption of various sensors, and the transmitting and receiving unit was
calculated and it is shown in Figure 24.
The transmitted sensor data was collected in the Google firebase cloud as shown in Figure 25
and it could be synchronized with the cloud messaging service and shared across the globe through
Sensors 2018, 18, 3318 20 of 23
the firebase cloud services.

Power consumption(W) High Power

Power consumption(W) Low Power
3 2.7

Power Consumption (W)

2.48 2.48
2.5

2
1.5
1.5
1.1
1
0.5
0.5

0
Raspberry Pi Sensing unit Transceiving Unit
IoT Hardware
Figure Power
24.24.
Figure consumption
Power ofof
consumption IoT system.
IoT system.
Sensors 2018, 18, 3318 19 of 22

Figure
Figure 25.
25. Google
Google fire
fire database.
database.

Water quality analysis

Water quality analysis waswas performed
performed using using the
the various
various sensors
sensors shown
shown in in Figure
Figure 26. 26. Water
Water
samples typical lake near to Ambattur, Chennai, India ◦ 0
samples were collected at a typical lake near to Ambattur, Chennai, India (13 06 27.9” N, 80 08 42.0” E)
were collected at a (13°06′27.9″ N, ◦ 0
80°08′42.0″ E)
extending
extending overover an
anarea
areaofofapproximately
approximately 1.57
1.57 kmkm 2 and has a total length of 6.06 km. A robotic arm
2 and has a total length of 6.06 km. A robotic arm with
with a water-sucking
a water-sucking pumppump was usedwastoused to collect
collect the water theand
water
it wasandstored
it wasinstored in a container.
a container. The complete The
complete sensor module was isolated to avoid interference of the signal,
sensor module was isolated to avoid interference of the signal, and their probes were immersed into and their probes were
immersed
the container.into
In the
ordercontainer. In order
to demonstrate to demonstrate
usefulness usefulness
of the developed of the developed
amphibious vehicle, water amphibious
sampling
vehicle, water sampling measurements were carried out at five different water
measurements were carried out at five different water locations across the lake. The real time data locations across the
lake. The real time data was transmitted for a period of time until saturation
was transmitted for a period of time until saturation occurred. The measured data was compared occurred. The measured
data
with was
theircompared
upper limitwith their
(Table 12)upper
basedlimit
on the(Table
Indian12)Standard
based on(IS) the10500
Indian Standard
water quality(IS) 10500 water
standards. It is
quality standards. It is adopted by the Bureau of Indian Standards approved
adopted by the Bureau of Indian Standards approved by the Drinking Water Sectional Committee and by the Drinking Water
Sectional
also the FoodCommittee and alsoDivision
and Agriculture the Food and Agriculture
Council. Division Council.
In situ measurement showed In thatsitu
all measurement
the measured
showed
data were almost equal as observed in Table 12. The maximum saturation time of 65 ms in maximum
that all the measured data were almost equal as observed in Table 12. The the case of
saturation
the turbidity time of 65was
sensor msobtained
in the caseandofa the turbidity
minimum sensor
of two was obtained
minutes interval wasand given
a minimum
between of each
two
minutes interval
sensing. The was given
collected water between each also
samples were sensing.
testedThe collected
in the water samples
lab environment and theweretestalso tested
results werein
the lab environment and the test results were in close agreement with the in
in close agreement with the in situ measurement having a 98% accuracy. It is evident from Table 12situ measurement having
athat
98%theaccuracy.
lake water It was
is evident
of poorfrom Table
quality and12it that the water
needed lake water was of
treatment poor quality
to improve and itquality.
the water needed
water treatment to improve the water quality.
showed that all the measured data were almost equal as observed in Table 12. The maximum
saturation time of 65 ms in the case of the turbidity sensor was obtained and a minimum of two
minutes interval was given between each sensing. The collected water samples were also tested in
the lab environment and the test results were in close agreement with the in situ measurement having
a 98%2018,
Sensors accuracy.
18, 3318 It is evident from Table 12 that the lake water was of poor quality and it needed
21 of 23
water treatment to improve the water quality.

(a) pH (b) Dissolved Oxygen

(c) Electrical Conductivity (d) Turbidity

Figure
Figure 26. Water quality
26. Water quality sensors.
sensors.
Table 12. Comparison of water quality with reference to IS 10500 standards.

Locations
Serial No. Sensors Maximum Allowable Limit
L1 L2 L3 L4 L5
(acceptable range = 6.5 to 8.5)
1 pH 8.12 7.98 7.83 8.00 7.01 >7.0 + = alkalinity
<7.0 − = acidity
2 Turbidity (NTU) 8.47 8.04 8.01 8.94 8.85 5.0
0–0.5 mS/cm Good
Electric conductivity
3 1.31 1.31 1.34 1.37 1.34 0.5–1.5 mS/cm Normal
(ms/cm)
>1.5 mS/cm High
4 Dissolved oxygen (mg/L) 8.75 8.75 9.04 8.82 8.61 Above 6 mg/L

10. Conclusions
An amphibious vehicle was developed for a mission endurance of 25 min while carrying a
payload of 7 kg. Design of the vehicle combined the functionalities of a multi-rotor UAV and a
hovercraft. Through engineering analysis and simulations, performance of the vehicle was evaluated
with reference to deformation, stresses, forward velocity, and stagnation pressure corresponding to
expected operational conditions. Appropriate selection of materials for obtaining superior strength
characteristics, motors, and propeller to generate sufficient thrust forces and considering a two DOF
robotic arm integrated with a water-sucking pump, a prototype was built and tested in an airborne
condition (open field) and also in a water body to evaluate the stability and response. The developed
amphibious system was able to collect water samples of 500 mL through actuating the suction pump
attached at the end-effector of the robotic arm. IoT-based water quality analysis revealed that within
11 ms, the 4G-LTE network transmitted the data to the ground station through firebase cloud services.
The developed IoT hardware unit consumed 7.58 W of power and each sensor saturation limit was
measured. The turbidity sensor took 65 ms to reach saturation and the pH required 26 ms to attain
saturation of the sensed data. Water quality analysis results suggested that, as per IS 10500 water
quality standards, the inspected lake water was impure and may not be suitable for drinking purposes.
Sensors 2018, 18, 3318 22 of 23

Author Contributions: B.E. developed the conceptual design of the amphibious vehicle, did the finite element
analysis, and prepared the entire manuscript. S.G. calculated the motor, propeller, and hovercraft system
parameters using empirical relations, design calculations, and computational fluid dynamics analysis. S.M. and
B.G. were involved in the prototype development, and testing of the amphibious vehicle and water quality
analysis. K.R. and J.S.C. formulated the IoT protocols and real time performance testing. S.W.P. and B.S. designed
the robotic arm for water sampling and water quality analysis using various sensor modules.
Funding: This research was supported by the NRF Korea-India Science and Technology Cooperation Expansion
Project (NRF—2017K1A3A1A57093906) and the DGIST R&D program of the Ministry of Science and ICT (18-IT-02,
20180463) and also supported by the DST—GITA (Ref: 2015RK0201103).
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Valavanis, K.P.; Vachtsevanos, G.J. Future of unmanned aviation. In Handbook of Unmanned Aerial Vehicles;
Springer: Dordrecht, The Netherlands, 2015; pp. 2993–3009.
2. Hassanalian, M.; Abdelkefi, A.B. Classifications, applications, and design challenges of drones: A review.
Prog. Aerosp. Sci. 2017, 91, 99–131. [CrossRef]
3. Prakash, N.U.; Vasantharaj, R.; Balasubramanian, E.; Bhushan, G.; Das, S.; Eqbal, F. Design, development
and analysis of air mycoflora using fixed wing unmanned aerial vehicle (UAV). J. Appl. Sci. Eng. 2014, 17,
1–8.
4. Sankarasrinivasan, S.; Balasubramanian, E.; Karthik, K.; Chandrasekar, U.; Gupta, R. Health monitoring of
civil structures with integrated UAV and image processing system. Procedia Comput. Sci. 2015, 54, 508–515.
[CrossRef]
5. Yang, L.J.; Esakki, B.; Chandrasekhar, U.; Hung, K.C.; Cheng, C.M. Practical flapping mechanisms for 20
cm-span micro air vehicles. Int. J. Micro Air Veh. 2015, 7, 181–202. [CrossRef]
6. Udayagiri, C.; Kulkarni, M.; Esakki, B.; Pakiriswamy, S.; Yang, L.J. Experimental Studies on 3D Printed Parts
for Rapid Prototyping of Micro Aerial Vehicles. J. Appl. Sci. Eng. 2016, 19, 17–22.
7. Chandrasekhar, U.; Yang, L.J.; Esakki, B.; Suryanarayanan, S.; Salunkhe, S. Rapid prototyping of flapping
mechanisms for monoplane and biplane ornithopter configurations. Int. J. Mod. Manuf. Technol. 2017, 9,
18–22.
8. Collins, K.A. A Concept of Unmanned Aerial Vehicles in Amphibious Operations. Ph.D. Thesis, Naval
Postgraduate School, Monterey, CA, USA, 1993.
9. Boxerbaum, A.S.; Werk, P.; Quinn, R.D.; Vaidyanathan, R. Design of an autonomous amphibious robot for surf
zone operation: Part I mechanical design for multi-mode mobility. In Proceedings of the 2005 IEEE/ASME
International Conference on Advanced Intelligent Mechatronics, Monterey, CA, USA, 24–28 July 2005; IEEE:
Piscataway, NJ, USA, 2005; pp. 1459–1464.
10. Yayla, M.; Sarsilmaz, S.B.; Mutlu, T.; Cosgun, V.; Kurtulus, B.; Kurtulus, D.F.; Tekinalp, O. Dynamic
Stability Flight Tests of Remote Sensing Measurement Capable Amphibious Unmanned Aerial Vehicle
(A-UAV). In Proceedings of the 7th Ankara International Aerospace Conference AIAC, Ankara, Türkiye,
11–13 September 2013.
11. Pisanich, G.; Morris, S. Fielding an amphibious UAV: Development, results, and lessons learned. In
Proceedings of the 21st Digital Avionics Systems Conference, Irvine, CA, USA, 27–31 October 2002; IEEE:
Piscataway, NJ, USA, 2002; Volume 2, p. 8C4.
12. Hasnan, K.; Ab Wahab, A. First Design and Testing of an Unmanned Three-mode Vehicle. Int. J. Adv. Sci.
Eng. Inf. Technol. 2012, 2, 13–20.
13. Frejek, M.; Nokleby, S. Design of a small-scale autonomous amphibious vehicle. In Proceedings of the
Canadian Conference on Electrical and Computer Engineering (CCECE 2008), Niagara Falls, ON, Canada,
4–7 May 2008; IEEE: Piscataway, NJ, USA, 2008; pp. 781–786.
14. Eaton, A.D.; Franson, M.A.H. Standard Methods for the Examination of Water & Wastewater; American Public
Health Association: New York, NY, USA, 2005.
15. Schaeffer, B.A.; Schaeffer, K.G.; Keith, D.; Lunetta, R.S.; Conmy, R.; Gould, R.W. Barriers to adopting satellite
remote sensing for water quality management. Int. J. Remote Sens. 2013, 34, 7534–7544. [CrossRef]
16. Gholizadeh, M.; Melesse, A.; Reddi, L. A comprehensive review on water quality parameters estimation
using remote sensing techniques. Sensors 2016, 16, 1298. [CrossRef] [PubMed]
Sensors 2018, 18, 3318 23 of 23

17. Peters, C.B.; Zhan, Y.; Schwartz, M.W.; Godoy, L.; Ballard, H.L. Trusting land to volunteers: How and why
land trusts involve volunteers in ecological monitoring. Biol. Conserv. 2017, 208, 48–54. [CrossRef]
18. Tyler, A.N.; Hunter, P.D.; Carvalho, L.; Codd, G.A.; Elliott, J.A.; Ferguson, C.A.; Hanley, N.D.; Hopkins, D.W.;
Maberly, S.C.; Mearns, K.J. Strategies for monitoring and managing mass populations of toxic cyanobacteria
in recreational waters: A multi-interdisciplinary approach. Environ. Health 2009, 8, S11. [CrossRef] [PubMed]
19. Karimanzira, D.; Jacobi, M.; Pfuetzenreuter, T.; Rauschenbach, T.; Eichhorn, M.; Taubert, R.; Ament, C. First
testing of an UAV mission planning and guidance system for water quality monitoring and fish behaviour
observation in net cage fish farming. Inf. Process. Agric. 2014, 1, 131–140.
20. Amyot, J.R. (Ed.) Hovercraft Technology, Economics and Applications; Elsevier: Amsterdam, The Netherlands,
2013; Volume 11.
21. Detweiler, C.; Griffin, B.; Roehr, H. Omni-directional hovercraft design as a foundation for MAV education.
In Proceedings of the 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS),
Vilamoura, Portugal, 7–12 October 2012; IEEE: Piscataway, NJ, USA, 2012; pp. 786–792.
22. Schleigh, J. Construction of a Hovercraft Model and Control of Its Motion. Ph.D. Thesis, University of
Maryland, College Park, MD, USA, 2006.
23. Rashid, M.Z.A.; Aras, M.S.M.; Kassim, M.A.; Ibrahim, Z.; Jamali, A. Dynamic Mathematical Modeling and
Simulation Study of Small Scale Autonomous Hovercraft. Int. J. Adv. Sci. Technol. 2012, 46, 95–114.
24. Fuller, S.B.; Murray, R.M. A hovercraft robot that uses insect-inspired visual autocorrelation for motion
control in a corridor. In Proceedings of the 2011 IEEE International Conference on Robotics and Biomimetics
(ROBIO), Phuket, Thailand, 7–11 December 2011; IEEE: Piscataway, NJ, USA, 2011; pp. 1474–1481.
25. Amiruddin, A.K.; Sapuan, S.M.; Jaafar, A.A. Development of a hovercraft prototype with an aluminium hull
base. Int. J. Phys. Sci. 2011, 6, 4185–4194.
26. Koko, M.I.A.A. Design of a Typical Multi-Role Vehicle Using Quad-Rotor Theory. Ph.D. Thesis, Sudan
University of Science and Technology, Khartoum, Sudan, 2014.
27. Estimation of Propeller Static Thrust. Available online: https://rcplanes.online (accessed on 1 July 2018).
28. T-Motors. Available online: http://store-en.tmotor.com (accessed on 1 July 2018).
29. Haider, A.; Sajjad, M. Structural design and non-linear modeling of a highly stable multi-rotor hovercraft.
Control Theory Inform. 2012, 2, 24–35.
30. Weerasinghe, R.; Monasor, M. Simulation and experimental analysis of hovering and flight of a quadrotor.
In Proceedings of the HEFAT 2017, Portorož, Slovenia, 17–19 July 2017.
31. Kuantama, E.; Craciun, D.; Tarca, R. Quadcopter Body Frame Model and Analysis. Ann. Univ. Oradea 2016,
71–74. [CrossRef]
32. Ozdemir, U.; Aktas, Y.O.; Vuruskan, A.; Dereli, Y.; Tarhan, A.F.; Demirbag, K.; Inalhan, G. Design of a
commercial hybrid VTOL UAV system. J. Intell. Robot. Syst. 2014, 74, 371–393. [CrossRef]

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).