unmanned aerial vehicle (uav)-assisted water sampling
TRANSCRIPT
Clemson UniversityTigerPrints
All Theses Theses
8-2016
Unmanned Aerial Vehicle (UAV)-Assisted WaterSamplingCengiz KoparanClemson University, [email protected]
Follow this and additional works at: https://tigerprints.clemson.edu/all_theses
This Thesis is brought to you for free and open access by the Theses at TigerPrints. It has been accepted for inclusion in All Theses by an authorizedadministrator of TigerPrints. For more information, please contact [email protected].
Recommended CitationKoparan, Cengiz, "Unmanned Aerial Vehicle (UAV)-Assisted Water Sampling" (2016). All Theses. 2461.https://tigerprints.clemson.edu/all_theses/2461
UNMANNED AERIAL VEHICLE (UAV)-ASSISTED WATER SAMPLING
A Thesis
Presented to
the Graduate School of
Clemson University
In Partial Fulfillment
of the Requirements for the Degree
Master of Science
Plant and Environmental Sciences
by
Cengiz Koparan
August 2016
Accepted by:
Dr. A. Bulent Koc, Committee Chair
Dr. Charles V Privette, III
Dr. Calvin B. Sawyer
Dr. Julia Sharp
ii
ABSTRACT
Water quality assessment programs require the collection of water samples for
physical, chemical, and bacteriological analysis. Lack of personnel, accessibility of water
bodies, and time constraints for water sampling, especially after natural disasters and
emergencies, are some of the challenges of water sampling. To overcome these
challenges, a water collection mechanism was developed and mounted on a multirotor
unmanned aerial vehicle (UAV) for autonomous water sampling from water bodies. The
payload capacity and endurance of the UAV (hexacopter) were verified using an indoor
test station. The hexacopter was equipped with floating foam, and the electronic
components were coated against water damage in case of landing on water due to
emergencies or water sampling. The system was able to collect water samples 48 times
out of 73 autonomous flight missions from a pond. The unsuccessful missions were
mainly due to the malfunctions of the servo motor used in water sampler’s triggering
mechanism. The servo motor for the mechanism was replaced to prevent the future
malfunctions. UAV-assisted autonomous water sampling is a promising method for
collection of water from water bodies. The system would be useful for collection of water
samples from large lakes or difficult to access water sources. The details of the developed
water sampling mechanism and the multirotor UAV, and experiment results are reported
in this thesis.
iii
ACKNOWLEDGMENTS
I would like to thank Dr. A. Bulent Koc for giving me the opportunity of being in
the graduate school and his limitless help, guidance, and support. I would also like to
extend my appreciation to my committee members; Dr. Charles Privette, Dr. Calvin
Sawyer, and Dr. Julia Sharp for their feedback, direction and assistance. I also would like
to thank Dan David and David Foltz for answering my questions related to thesis
formatting and editing. I thank Kevin Gibson for his help during data collection.
iv
TABLE OF CONTENTS
Page
ABSTRACT ......................................................................................................................... i
ACKNOWLEDGMENTS ................................................................................................. iii
TABLE OF CONTENTS ................................................................................................... iv
LIST OF TABLES ............................................................................................................ vii
LIST OF FIGURES ......................................................................................................... viii
CHAPTER
1. INTRODUCTION ...................................................................................................... 1
Importance of Water Quality Monitoring ........................................................................... 1
Limitations of Water Sampling ........................................................................................... 4
Possibility of a UAV for Water Sampling .......................................................................... 5
2. LITERATURE REVIEW ........................................................................................... 7
Fundamental Reasons for the Water Quality Research ...................................................... 7
Current Water Quality Monitoring Methods ...................................................................... 9
Precautions Regarding Compliance with FAA Regulations ............................................. 11
Objectives ......................................................................................................................... 13
3. MATERIALS AND METHODS .............................................................................. 14
Development of a UAV .................................................................................................... 14
Hexacopter Components ................................................................................................... 15
Flight Performance Station ............................................................................................... 17
v
Table of Contents (Continued)
Page
Dynamic Analysis of the Hexacopter ............................................................................... 19
Water Sampler Subsystem ................................................................................................ 23
Water Sampler Performance Tests.................................................................................... 28
Experiment Location ......................................................................................................... 30
Autonomous Flight Missions for Water Sampling ........................................................... 31
Measurement Procedures of Water Quality Indices ......................................................... 33
4. RESULTS AND DISCUSSIONS ............................................................................. 36
Payload Capacity of the UAV........................................................................................... 36
Theoretical Endurance of the UAV .................................................................................. 38
Experimental Results of Water Sampler Performance ..................................................... 39
Estimation of Water Sampling Depth ............................................................................... 40
Chloride Accuracy Check ................................................................................................. 43
Performance of Water Sampler During Autonomous Water Sampling ............................ 44
Consistency Between Manual and UAV-Assisted Sampling ........................................... 48
5. CONCLUSIONS....................................................................................................... 52
6. FUTURE WORK ...................................................................................................... 53
APPENDICES .................................................................................................................. 55
Appendix A. LabVIEW® Programs used for load cell calibration and load measurement.
........................................................................................................................................... 56
Appendix B. Microcontroller programing code for automatic messenger release. .......... 57
vi
Table of Contents (Continued)
Page
Appendix C. Autonomous flight mission path and water sampling commands. .............. 58
Appendix D. Field notebook of water sampling and on-site analysis .............................. 59
Appendix E. SAS output for sampling depth analysis ...................................................... 60
Appendix F. Pre-flight check list before autonomous water sampling. ............................ 62
Appendix G. Autonomous water sampling locations and boundaries of sampling site. .. 64
Appendix H. SAS Output for water sampling consistency analysis. ................................ 65
vii
LIST OF TABLES
Table Page
Table 1. Components of the hexacopter with their specifications. ................................... 16
Table 2. Load cell calibration data. ................................................................................... 18
Table 3. Total weight of the hexacopter with payload. ..................................................... 37
Table 4. Average thrust and endurance at different throttle levels. .................................. 38
Table 5. Number of successful and failed sampling flights during the initial tests. ......... 40
Table 6. Autonomous flight information for altitude range experiments. ........................ 42
Table 7. Least Squares Means results from SAS®. .......................................................... 43
Table 8. Chloride accuracy check and percent recovery. ................................................. 44
viii
LIST OF FIGURES
Figure Page
Figure 1. Locations of impaired waterbodies listed in EPA 303(d) list in South Carolina
(Wilson 2015). .................................................................................................................... 8
Figure 2. Close-up images of the water sampling system: a) hexacopter, b) ultrasound
range sensor, c) microcontroller, d) servo motor, e) water sampler, f) messenger, g)
floating attachments, and h) hexacopter with all of the attachments. ............................... 15
Figure 3. Flight performance test station for indoor experiments with load cell, signal
conditioner and data acquisition board. ............................................................................ 17
Figure 4. Load cell calibration graph. ............................................................................... 19
Figure 5. Kinematic scheme of the hexacopter. ................................................................ 20
Figure 6. 3D design from SolidWorks (left) and the 3D printed water sampler (right). .. 24
Figure 7. Location of the Lamaster Pond at Clemson University campus. ...................... 30
Figure 8. Schematic showing the operation of the UAV-assisted water sampling. .......... 32
Figure 9. UAV indoor flight performance test results. ..................................................... 37
Figure 10. Change of thrust with endurance of the UAV at 50% throttle level. .............. 38
Figure 11. UAV with water sampler hovering at water sampling altitude during an
autonomous flight (left) and autonomous flight path for aerial water sampling (right). .. 41
Figure 12. Barometer altitude measurements during 18 seconds of water sampling from
the first autonomous flight. ............................................................................................... 42
Figure 13. Chloride accuracy chart. .................................................................................. 44
ix
List of Figures (Continued)
Page
Figure 14. Aerial image which was taken at 120 m altitude represents the sampling
points. ................................................................................................................................ 46
Figure 15. Water constituent measurements from hand sampling and UAV-assisted
sampling. The points represent the average of three replicate measurements for
temperature (a), pH (b), DO (c), EC (d), and Chloride (e). .............................................. 47
Figure 16. Interaction plots between hand and UAV-assisted water constituent
measurements for temperature (a), pH (b), DO (c), EC (d), and chloride (e). .................. 51
1
CHAPTER
1. INTRODUCTION
Importance of Water Quality Monitoring
Water quality monitoring is necessary for many purposes. One of these purposes
is to characterize water and identify changes or trends in water quality over time. Water
quality monitoring also can be made for identification of specific existing or emerging
water quality problems. For example, pollutants carried by stormwater may include
bacteria, nutrients, litter, sediment, oils, and heavy metals (Thomas et al., 2001). Through
monitoring, information can be gathered to implement specific pollution prevention and
remediation programs. In addition, water monitoring allows checking whether the
program goals are in compliance with the water pollution regulations. Water quality
monitoring can also be useful for rapid response to flooding events after natural disasters.
Quality of water samples is one of the most important factors that effects the
analysis results. Depending on the type of analysis, the time frame used to deliver water
samples to a laboratory is essential. Ideally, an infrastructure should be established to
enable all of the water samples to be returned to a central or regional laboratory within a
few hours of sample collection (Bartram and Balance, 1996). However, this scenario
depends on the quality of the road system and the reliability of motorized vehicles for
sampling officers. Yet, these amenities are not readily available in many regions and
countries.
2
There are various methods available for water sampling. The methods used for
surface-water sampling often depend on the flow characteristics of surface water. In
addition, safety of field personnel, suitability of the equipment for analysis, field
measurement profiles, temporal and spatial heterogeneity, ecological characteristics,
weather conditions, fluvial-sediment transport, and the point and nonpoint sources of
contaminations are other considerations for choosing an appropriate water sampling
method (Myers, 2006). These considerations need to be made in terms of preventing
biases induced in water property measurement. Biases caused by the water collection
process need to be minimized by selecting a good sampling site and sampling technique.
Each sample should represent the intended environmental condition at the time of
sampling.
Streams receive inflow from many sources and entry routes. Streams receive
pollutant entry from two different routes known as point sources and nonpoint sources
(Liu et al., 2016). Drainage channels, outlets from industrial plants, and sewers are some
of the point sources of inflow. Whereas, water pollutant entry from nonpoint sources
occurs only after rainfall or emergency overflows in a short period of time. For example,
water that runs over impervious surfaces such as roadways, rooftops, parking lots and
sidewalks picks up pollutants along the way and carries them directly to lakes, rivers and
estuaries (Ma et al., 2016). Water that heats up on parking lots and roadways also can
lead to warmer than normal water entering nearby waterways which would cause increase
in temperature of the surface water. This runoff also called stormwater is generated by
precipitation, melting snow and irrigation water that runs off the land (Lim and Lu,
3
2016). Monitoring water quality after a storm event would require a rapid water quality
assessment technique to determine the sources of contaminations otherwise it would be
difficult to identify the source accurately.
In addition to rapid water sampling after a storm event, regular water sampling is
also necessary to identify the entry points of pollutants into the surface water. For
example, nutrient leaching from farm fields or pasture lands to surface water would cause
algal blooms. Growth of dense algal blooms causes discoloration in water bodies. Among
algal blooms, blue-green algae has the generic potential to produce toxins (van der
Merwe, 2015) which are harmful to humans and animals and cause death of livestock
(McGowan et al., 2016). Water monitoring programs and emergency plans can help
reduce the impacts of algal blooms. Regular water sampling and quality analysis provide
an opportunity to notify the public to obtain drinking water from alternative sources in
case of possible harmful algal blooms.
Receiving water pollutant entry from nonpoint sources occurs only after rainfall
or emergency overflows in a short period of time. These entries can be monitored with
event-controlled water samplers (Neumann et al., 2003). Therefore, a rapid water quality
assessment technique is needed to understand nonpoint source contaminations for
improved water quality monitoring.
4
Limitations of Water Sampling
Limitations of water sampling include the velocity and depth of stream waters.
The velocity and the depth of sampling location affect the quality of the water sampling.
In order to collect appropriate water samples safely, flowing water should not be waded
into if the measurement depth (m) of stream multiplied by its velocity (m/s) equals 0.93
m2/s (10 ft2/s) or greater (Wilde et al., 2003).
On the contrary, water quality indices have traditionally been measured with in-
situ sampling. Therefore, measurements retrieved by a sample represent particular
location and time (Madrid and Zayas, 2007). This type of measurement may cause
problems when the time and space scales are not reached (Herring et al., 1990). In
addition, in-situ sampling may provide inaccurate measurements within its surrounding
on different days (Erkkilä and Kalliola, 2004). Even though measurements may be highly
accurate for a specific spot, the data may not correspond well with a sample taken at a
nearby location (Luhtala and Tolvanen, 2016). Additionally, sampling location is an
important factor for effectiveness of water quality monitoring. The sampling site should
represent the environment under study. A large number of sites would provide more
information about water quality but cost, accessibility, and time are inhibiting factors
(Madrid and Zayas, 2007). Temporal resolution of water samples is another important
factor that affects representativeness of the surrounding environment. For example,
runoff from periodic agricultural pesticide applications leads to spatial and temporal
variations in the physico-chemical characteristics of water (Warren et al., 2003).
5
Possibility of a UAV for Water Sampling
UAVs are used for different environmental monitoring tasks including thermal
infrared imaging of geothermal environments (Nishar et al., 2016), coastal surveying for
topographic mapping and measurement (Turner et al., 2016), and marine wild life
monitoring (Fortuna et al., 2013) for detecting and tracking hammerhead sharks. UAVs
are suitable for remote sensing applications because of their capability of carrying sensors
and ability to monitor large areas in a relatively short period of time (Nilssen et al.,
2015). These platforms can be influenced by various external and internal factors
including change in wind speed, surrounding obstacles such as trees, and safety issues
related to operational difficulties (Patterson et al., 2014). These factors can cause
problems during water sampling; therefore, they may cause challenges.
The challenges of obtaining water samples via a UAV needs to be addressed
before launching a sampling mission over a water body. It is essential to confirm that the
key water chemicals are not biased by using a UAV-mounted mechanism. Materials that
will be used to construct the water-sampling equipment must be defined for the purpose
of water sampling. Some materials tend to react with water, therefore water sampling
equipment may induce a bias in measurements from organic or inorganic analysis (Wilde
et al., 2003). In order to eliminate and reduce chemical reactions, proper materials must
be selected for the type of analysis.
The UAV that is used for water quality monitoring needs to be able to carry the
additional weight of the sampler mechanism, collected water, and additional parts that
allows the UAV to float on the water surface in case of an emergency landing. These
6
attachments are called as payload (Shan, 2016). The UAV has to produce enough thrust
for lifting its payload (Barve and Patel, 2014). Thrust is one of the basic motions of a
multirotor generated by its propellers. Developed thrust force is directly proportional to
the rotating torque of motors (Atlam and Kolhe, 2013). Therefore, motor size is an
important factor for producing required thrust for the UAV. There are various types of
motors in the market that are used for UAVs. Each of these motors have different current
consumption, and propeller size needed to create enough thrust to lift the vehicle and its
payload. The correct size of the motors must be selected to produce thrust for the aircraft,
and keep it floating in the air (hovering) during water sampling.
7
2. LITERATURE REVIEW
Fundamental Reasons for the Water Quality Research
UAV-assisted rapid water sampling method would provide quick and accurate
water quality data. Water sampling after contamination from nonpoint sources can be
made with ease considering the difficulty of determining the nonpoint sources with
stationary sensors. UAV-assisted water sampling will help to determine the sources of
emerging water runoffs and this information would be used by Environmental Protection
Agency (EPA) and other responsible organizations to create further monitoring programs.
EPA is responsible for setting up water quality standards and mandates that the
303(d) List of impaired waters be reviewed every two years (Copeland, 2012). This list
names waterbodies that do not meet water quality standards; the locations of the impaired
water bodies in South Carolina are shown in Figure 1. The purpose of the list is to
identify impaired waters to help describe the sources of impairment so that the necessary
actions can be taken to improve water quality.
8
Figure 1. Locations of impaired waterbodies listed in EPA 303(d) list in South Carolina (Wilson, 2015).
South Carolina Department of Health and Environmental Control (SCDHEC)
considers all available water quality data to be used for developing the 303(d) List. This
water quality data can be obtained from individuals who voluntarily submit their results.
These volunteers are required to submit a Quality Assurance Project Plan (QAPP) for
approval prior to water sampling (SCDHEC, 2015). Even though volunteer water quality
programs provide help and guidance for water quality monitoring data collection in South
Carolina (Addy et al., 2010), there are surface waters that are not linked to other larger
surface waters so that access via boat may not be possible. Therefore, water sampling
from these surface waters can be a challenging task.
Additionally, volunteer water quality monitoring programs address sampling
shortfall from watersheds. A volunteer program called Adapt-a-Stream (AAS), developed
9
by the Georgia Environmental Protection Division, addresses sampling shortfall and
monitors water quality issues for Twelve Mile Creek located in Pickens County, SC. It
was projected that E. coli levels increased significantly during high-flow discharge due to
storm events among sites located near Twelve Mile Creek Blueway (Nation and Johnson,
2015).
In practice, cost, time, and labor limit the sampling interval and the spatial
resolution that can be attained from surface waters. A UAV-assisted water sampling
system can be used to collect water samples to estimate water quality and help improve
the problem identification process during or after a storm event. High-resolution water
quality measurements from surface waters are essential to identify these problems
(Dlamini et al., 2016). Mobile, low-cost, and autonomous water sampling systems could
provide high resolution water quality data.
Current Water Quality Monitoring Methods
Different water quality monitoring methods are currently in use. One of these
methods is satellite-based turbidity sensor stations. These satellite-based turbidity sensors
in river streams are available to measure the sediment amount in water (SCWRRI, 2004).
These sensors have limited speed of data transfer that prevents policy makers from acting
quickly to make daily decisions in order to protect water quality. In addition, data sent by
satellite-based turbidity sensors may not be accurate and reliable due to the signals
interactions with the atmosphere. Atmospheric effects such as attenuation caused by
10
cloud, rain, and atmospheric gases may affect the quality of data that the sensors measure
(Campbell and McCandless Jr., 1996).
Another method for water quality monitoring is using autonomous vehicles for
water sampling or in-situ measurement in waterbodies. One of the examples for these
autonomous vehicles is Autonomous Underwater Vehicle (AUV). AUVs have been used
for water quality monitoring in recent years. These vehicles autonomously monitor the
water quality while they travel under water. However, water quality monitoring with
AUVs can be challenging. The major challenge for this type of underwater vehicle is that
when it is underwater the GPS system cannot be used for accurate positioning
(Karimanzira et al., 2014). Because of this reason, the AUV has to be equipped with
additional navigational systems or requires some acoustic localization.
Another example for autonomous vehicles which are used for water quality
monitoring is Autonomous Surface Vehicle (ASV). An unmanned airboat was used for
automatic water quality indices measurements in a recent research project. The vehicle
automatically navigates to predefined sampling points and measures pH, DO, EC,
turbidity, temperature, sensor depth, water depth, chlorophyll-a concentration, and NO3.
Disadvantages of automatic sampling with a boat were the operational difficulties due to
swaying from side to side and uncertain engine-control frequencies (Kaizu et al., 2011).
Unlike the above methods, an aircraft or a UAV can be used to collect aerial
images of waterbody of interest. Aerial images that are taken remotely with a UAV can
help visualize the disturbance in the water and provide enhanced spatial water quality
monitoring data. An aerial survey with a high resolution camera attached to a UAV can
11
also show the locations of water contamination. Aerial images taken from a UAV can
provide the coordinates of contaminations which can help determine more accurate water
sampling locations. These sampling points would increase the accuracy and precision of
water quality data. Necessary actions regarding public safety can be taken with the help
of rapid water quality data.
Precautions Regarding Compliance with FAA Regulations
The UAVs need to be registered under Federal Aviation Administration (FAA) to
ensure compliance with regulations. A copy of registration needs to be carried along with
the UAV during experiments and a certificate number should be attached on the UAV.
According to FAA, UAV systems are divided into three different groups based on their
usage or interest area. If a UAV is used for public operations, it falls in the governmental
UAV operations group. If a UAV is used for civil operations, it falls in the non-
governmental UAV operations group. If a UAV is used for hobby or recreational
purposes, it falls in the model aircraft operations group (FAA, 2016).
Public UAV activities are limited by federal laws to certain operations within the
U.S. airspace and require a Certificate of Waiver or Authorization (COA). A public
aircraft is defined as an aircraft owned by the government and operated by any person for
purposes related to crew training, equipment development or demonstration. The COA
permits public agencies and organizations to operate an aircraft for different purposes in
particular area for a limited time. This permit can last up to 2 years in most cases. Some
of the public UAV operations include law enforcement, firefighting, and border control,
12
as well as search and rescue. The typical COA application approval process is within 120
business days (FAA 2016).
If a UAV is not used for public operations, then it’s considered as non-
governmental UAV operation and it must be operated in accordance with all of the FAA
regulations. Civil operations also require authorization. There are two methods of gaining
an authorization for non-governmental UAV: the first one is Section 333 Exemption and
the second one is Special Airworthiness Certificate (SAC) (FAA, 2016). Basically,
Section 333 Exemption provides a grant for COA, but this exemption may be used for
commercial operations in controlled environments. A petition for exemption can be found
at the FAA’s website (faa.gov), including instructions for how to apply.
On the other hand, SAC falls in the experimental category and is required for
research and development, crew training, and market survey. Only a SAC holder can
operate a UAV for any of the categories listed above. FAA inspectors can issue a SAC
and special flight permit using FAA order 8130.34. If the FAA inspector decides that the
research or project requires additional safety precautions, an additional Special
Airworthiness Certificate in the Experimental Category will be issued. This additional
certificate will include operation limitations for intended UAV for that specific operation
(FAA, 2016).
Model UAV operations include hobby and recreational use and do not require any
certifications unless safety guidelines are followed. These safety guidelines are created by
FAA and the security of national airspace. One of the most important safety guidelines of
model UAV operations is the maximum 500 ft (152 m) altitude restriction. The aircraft
13
must be kept within the line of sight during the operation and a distance of 5 miles (8 km)
from the nearest airport must be maintained unless the airport and control tower are
informed before the operation. The weight limit of the aircraft is 55 lbs (25 kg) and
flights near stadiums or crowds are prohibited. The operator is always responsible for
his/her actions during the flight and an operator could be fined for endangering people or
other aircrafts.
Objectives
The goal of this research is to develop an Unmanned Aerial Vehicle (UAV)-assisted
rapid water sampling system for surface water monitoring. The specific objectives to
reach this goal are:
- To develop a UAV-assisted water sampling system that automatically collects 130
ml of water sample per flight.
- To integrate a subsystem which includes a sensor for altitude approximation to
enable automatic sampling at low altitudes over water surface.
- To test the flight performance of the system indoors in a flight performance
station.
- To confirm that the UAV-mounted water sampling mechanism does not bias the
water quality indices of Dissolved Oxygen (DO) content, Temperature, Electrical
Conductivity (EC), and Chloride level with outdoor experiments.
14
3. MATERIALS AND METHODS
Development of a UAV
A hexacopter was developed for this research. The control of the hexacopter was
accomplished with a radio controller (RC) (Turnigy 9X, Hextronik, China) or a flight
controller software. The main components of the hexacopter were a frame, motors,
propellers, electronic speed controllers, flight controller (Pixhawk, 3DR, USA), GPS
receiver (Ublox, 3DR, USA), telemetry radio transmitter (3DR radio, 3DR, USA), and a
power supply (Venom, USA). The payload mounted at the bottom side of the frame
contained a microcontroller (ATmega328P, San Jose, CA), ultrasound range sensor
(MB1040, MaxBotix, USA), servo motor, water capturing system, messenger to trigger
water capturing mechanism, and floating attachments (Figure 2). A ground control station
contained a computer which communicated with the UAV via a telemetry radio
transmitter. The ground station gathered information about the flight condition including
battery level, altitude, location, flight mode situation, and GPS connection status during a
flight mission. All electronic circuits in the system were waterproofed using a corrosion
prevention spray (90102 Anti corrosion, Corrosion-X, USA) to prevent water damage to
the system components.
15
Figure 2. Close-up images of the water sampling system: a) hexacopter, b) ultrasound range sensor, c)
microcontroller, d) servo motor, e) water sampler, f) messenger, g) floating attachments, and h) hexacopter
with all of the attachments.
Hexacopter Components
A stabilized control of a UAV is essential when flying above the water surface.
Multiple considerations must be evaluated when choosing electronic parts for these
vehicles; the electronic parts were chosen depending on the desired payload and flight
time as well as their compatibility (Gupta et al., 2014). The total payload capacity of a
UAV depends on the thrust that it can produce. Therefore, water sampling capacity was
limited. Indoor experiments were conducted to determine the flight duration of the UAV
with the water sampler. The specifications of the UAV components used in this research
are shown in Table 1.
16
Table 1. Components of the hexacopter with their specifications.
Part Name Model/Number Specifications Function Quantity Weight (g)
Motors T-Motor/MN3110 KV700 Main actuator 6 480g
ESCs Arris SimonK 30A Speed Control 6 150g
Propellers Andoer/1045 Carbon Fiber Propelling 6 51g
Microcontroller Pixhawk Open hardware Flight Control 1 182g
Battery (Li-Po) Venom 14.8V 25C 4S 5000mAh Power Supply 1 542g
Power Module 3DR-XT60 BEC 3A 5.3V 7S Power distribution 1 28g
RC receiver 9X 8Ch Mode2 2.4Ghz Receive commend 1 8.5g
Radio receiver 3DR 433Mhz Radio commend 1 8.5g
GPS 3DR-0003 Compass 5Hz GPS antenna 1 45g
Servo SG90 4.8V (1kg/cm) Triggers sampler 1 9g
Frame F550 PCB wiring Holds parts 1 678g
Floating parts N/A 753 cm3 Provides floating 12 151g
TOTAL 2333g
The number of cells in the battery must be considered when choosing motors and
electronic speed controllers (ESC) for the hexacopter. The size of the motors are limited
by the current allowance provided by the ESC. In order to provide enough lift and thrust
for the desired payload, six T-type brushless DC Motors were selected in compliance
with 10” propellers. MN3110 (T-Motor, USA) model motors can operate at 700 rpm/V
and can work with up to a 6-cell battery. In order to provide the required power and
control the motors’ rotational speed, six ESCs (Arrishobby, China) with 30A allowance
were used. The system was powered by a high discharge, light weight 14.8V Li-Po 6-cell
battery (Venom, USA). The Li-Po battery had a 25C discharge rate with 5000 mAh
capacity. The UAV was also controlled by a Turnigy 9X (Hextronik, China) model
2.4GHz radio controller.
17
Flight Performance Station
An indoor flight performance test station was built to verify the payload capacity
and the endurance of the battery at different throttle levels (Figure 3). The station
provided a safe housing with four wooden columns for the UAV testing. The hexacopter
once placed in the columns moved freely in the vertical direction. A load cell (MR01,
Honeywell, USA) was placed under the top cover of the station to measure the force
while the hexacopter was hovering in the upward direction. NI-USB-6009 (National
Instruments, USA) data acquisition board was used to transfer the load cell measurements
to a computer.
Figure 3. Flight performance test station for indoor experiments with load cell, signal conditioner and
data acquisition board.
Load Cell Calibration
NI-LabVIEW® software was used to calibrate the load cell and record the data
for thrust measurements. LabVIEW® provides an environment for custom applications
which interacts with real-world data. The programming language used in LabVIEW® is
18
known as G-language. G-language is a graphical dataflow language where functions
operate on the available data (National Instruments, 2013).
Before attempting to use the load cell to measure the thrust values, the load cell
was calibrated by converting real-time load values (g) into electrical signals (mV). The
load cell is a transducer which creates an electrical signal whose magnitude is directly
proportional to the force being measured (VPG Transducers, 2015). Capacity of the load
cell was 50lbs with an output voltage range of 0-10V. In order to calibrate the load cell,
the applied known weights were converted into mV values and transferred to the
computer. A trend line between the measured mV values and known masses was
developed. The linear trend line equation was then used for determining the unknown
forces acting on the load cell. Table 2 shows the load cell calibration data.
Table 2. Load cell calibration data.
Number Known load (g) Measured value (mV)
1 0 45.6
2 4150 322
3 5185 390.9
4 6862 502.3
5 11505 809
6 45353 3213
To determine the amount of thrust, the data measurements were used to create a
graph (Figure 4). The resulting regression equation was used in LabVIEW® to measure
the loads when the UAV pushed on the load cell (Appendix A). These load
measurements indicated the amount of thrust which UAV applied during indoor
experiments.
19
Figure 4. Load cell calibration graph.
Dynamic Analysis of the Hexacopter
Mathematical model of a hexacopter was derived in accordance with Newton’s
law of motion. All of the acting forces on the hexacopter need to be considered in order
to set up a dynamic model (Baránek and Šolc, 2012). Acting forces on the hexacopter are
mass, acceleration, gravity, propeller thrust, drag force, and disturbance (Figure 5).
Propeller thrust force on the UAV can be derived from equations to follow (Gupta and
Jha, 2014). As shown in the kinematic scheme of the hexacopter, the origin of reference
frame (Xr, Yr, Zr) is placed at the center of gravity (COG).
y = 0.0701x - 18.735
0
500
1000
1500
2000
2500
3000
3500
0 10000 20000 30000 40000 50000M
easu
red
volt
ag
e (m
V)
Applied Load (g)
mV
20
Figure 5. Kinematic scheme of the hexacopter.
Ti represents the positive direction of thrusts, Mi represents the positive direction
of reactive torque of each propeller, and ωi represents the positive direction of rotation.
The force generated by the UAV were calculated using Equation 1.
z
y
x
mF total
int (1)
Where,
Fint = Interactive forces acting on aircraft (N)
mtotal = Total mass of aircraft (kg)
ẍ = Acceleration in the X-direction (m/s)
ӱ = Acceleration in the Y-direction (m/s)
�̈� = Acceleration in the Z-direction (m/s)
21
Gravity pulls down the hexacopter in the Z-direction, downward direction in the
reference frame of the hexacopter. Gravity force on the hexacopter in the Z-direction was
calculated using Equation 2:
g
RMF NBtotalg 0
0
(2)
Where,
Fg = Gravity force acting in the Z-direction (N)
Mtotal = Total mass of aircraft (kg)
RNB = Normal reaction force due to gravity (N)
g = Gravity (m/s2)
The thrust from the propellers is always assumed to be in the positive Z-direction
in the frame body. According to Gupta et al. (2014), it was assumed that there was no
thrust in the X- and Y-directions (Xr and Yr). Total thrust produced by propellers was
calculated by using Equation 3:
M6M5M4M3M2M1
thrust
FFFFFF
0
0
F (3)
Where,
Fthrust = Propeller thrust (N)
FMi = Downward forces from individual motors (N)
Drag force is generated by the friction in the air, and acts opposite to the relative motion
of the aircraft (Xiong and Zheng, 2015). Drag force was calculated by using Equation 4:
22
Dwdrag CSVF 2
2
1 (4)
Where,
Fdrag = Drag force (N)
ρ = Density of the fluid (g/mL)
V = Speed of the object relative to the fluid (m/s)
Sw = Cross sectional area (m2)
CD = Drag coefficient
Outdoor environment for the UAV operations may contain external forces which
can cause disturbance. Outdoor disturbances are caused by external forces include wind
and gust (Doshi et al., 2015). These external forces might cause unstable flight or even
results for failure of mission and accident during landing of the UAV (Ambati and Padhi,
2016)Total force that affects the UAV was calculated by using Equation 5:
dragedisturbancthrustgravitytotal FFFFF (5)
Where,
Ftotal = Total forces action on the UAV (N)
Fgravitiy = Gravity force (N)
Fthrust = Thrust force (N)
Fdisturbance = Disturbance force (N)
Fdrag = Drag force (N)
The agility could be an important issue for multirotors when they are scaled up in
outdoor environments (Segui-Gasco et al., 2014). If the aircraft is scaled up, it requires
23
larger control moment to allow necessary angular acceleration on the airframe (Mahony
et al., 2012). As the weight increases, the propeller size must also increase. Therefore,
allowable angular acceleration is reduced and the frequency bandwidth of the aircraft
control is decreased (Cutler et al., 2011). Decreased frequency bandwidth limits the
response time for position control. Carbon fiber propellers with 10 in length and 4.5 in
pitch were selected in order to reduce agility and provide enough angular acceleration
while producing required thrust. On the other hand, if any of the motors fails during a
flight, the hexacopter would be de-stabilized (Lo et al., 2013) but it can still fly. This
would allow some time for safe landing and prevent a possible crash.
Water Sampler Subsystem
Water Sampling Equipment
A thief style water sampling equipment (Wilde et al., 2003) was designed with
SolidWorks® (Figure 6). Designed parts were 3D printed with a laser printer using
Selective Laser Sintering (SLS) technique (3D Printing, 2016). The water holding
capacity of the water sampling equipment was designed as 130 ml. The length and
diameter of the sampling tube were 15.24 cm and 3.81 cm, respectively. The water
sampler consists of a container, container covers, a release button, and a plastic string.
The container is where the water sample is stored. The plastic string closes the two open
ends rapidly once the release button is pushed. The release button is pushed when the
messenger is released at the bottom of the UAV. Releasing the messenger and triggering
the push button actions were achieved with the microcontroller.
24
Figure 6. 3D design from SolidWorks (left) and the 3D printed water sampler (right).
The main goal of the subsystem was to control the sampler with the
microcontroller. The microcontroller was equipped with an ultrasonic range sensor for
altitude approximation and a servo motor for triggering the water sampler. Sun-Founder
HC-SR04 (Clayton Technologies, Malaysia) model ultrasound sensor was used to acquire
distance information in order to control the sampling system. HC-SR04 has 40 Hz of
working frequency and 5V of working current according to manufacturer’s manual. The
working range of the ultrasonic sensor is between 2 cm and 4 m.
Working principals of the ultrasonic ranger is based on the soundwave emission
and transmission through the air. The ultrasonic sensor emits soundwaves into a specific
direction. If there is an object in that direction within the range, soundwaves bounce back
from that object to the sensor. The microcontroller measures the time that it takes for the
soundwaves to travel and calculates the distance between the object and the sensor
(Rockwell Automation, 2016). The ultimate purpose of ultrasound sensor was to estimate
25
the altitude over water surface and send range readings to the microcontroller. Based on
the sensor readings, the microcontroller would trigger the messenger. Messenger would
fall down by the gravity along the rope and hit the button on top of the water sampler and
release the end caps to enclose the water.
Microcontroller for Sampler Control
The microcontroller was programmed to test the water sampling mechanism
indoors. The microcontroller was programmed (Appendix B) in a way that it would
control the servo when a pre-determined altitude was reached. This pre-determined
altitude was the distance between the ultrasound sensor and the water surface. This
distance was kept to provide safe hovering for the UAV. Sampling altitude was set to 1m
during indoor tests in the flight performance station.
Ultrasound waves travel through the air at the speed of sound. The
microcontroller commends ultrasound sensor to produce a short 10µs pulse at eight cycle
bursts of ultrasound at 40 kHz and raise its echo. After soundwaves travel and bounce
back from the object, the microcontroller calculates the range through the time interval
between sending a trigger signal to generate soundwaves and receiving echo signal
(Clayton Technologies, 2013).
Range was calculated with microcontroller by using Equation 6:
2* VelocityTimeLevelHighRange (6)
Where,
Range = Distance from an object (m)
26
High Level Time = Time required for sound beams to travel at high pulse
(s)
Velocity = Speed of sound in air (340 m/s)
The ultrasound sensor constantly sends these sound beams to obtain continuous
range measurements. The microcontroller was programmed with a while loop so that the
range measurements would not affect the servo motor until the sampling altitude was
reached. Otherwise, the sampler would be triggered before the UAV launch because the
water sampling altitude was set to 1m. The loop code in the microcontroller allows servo
motor to wait until the UAV launches, navigates to the sampling point, and descends to
the water sampling altitude. The water sampler is triggered only after the ultrasound
sensor measures altitude below 1m for the second time. The UAV has to launch and
ascend to navigation altitude (10m). When the UAV descends to the sampling altitude,
the ultrasound sensor measures the distance of 1m and the microcontroller ends the loop,
and sends a command to release the messenger.
Difficulties with Sampler Control Using an Ultrasound Sensor
Due to an error in ultrasonic range measurements during flights, it was decided
that the ultrasound sensor assisted sampling control may not be feasible for this system.
The source of this type of error in an ultrasonic measurement was due to the variability of
sound speed in the transmission path between the sensor and the target. In addition, the
amplitude of echo signals might change by large amounts from pulse to pulse due to the
27
variations in the speed of sound in air. This might be caused by air turbulence or target
movements (Massa, 1999). Since this is a multirotor platform and it flies by moving air
downwards with propellers, it is most likely that an air turbulence is created between
ultrasonic sensor and the target. This turbulence could cause an error in acquiring correct
measurements; therefore, the microcontroller could receive uncertain measurements.
Because of this, the microcontroller might send a signal to enable the servo to release the
messenger before the UAV descends to the desired sampling altitude. In order to correct
this error and be able to use the ultrasound sensor as a reliable water sampler controlling
sensor, further experiments for filtering and signal processing was needed. Another issue
with the use of ultrasound range sensor was the interference of sampler with the range
measurements. The water sampler was hanging downward from the UAV with a rope.
During a flight mission, depending on the weather conditions and the UAV operation, the
sampler was interfering with range measurements by causing a surface to reflect the
soundwaves. The reflected soundwaves were making the range sensor to measure a range
which would be within the preset range. In order to eliminate this problem, the ultrasound
ranger was mounted 70cm away from the center of the aircraft (Figure 8). This helped the
ultrasound ranger to avoid the sampler from interfering with the range measurements and
decreased the risks of a possible interference, but it wasn’t reliable all of the time.
Because of the oscillation of the water sampler, the soundwaves were still getting blocked
occasionally during test flights; thus, causing the messenger to be released before the
desired condition.
28
Using flight controller to control water sampler subsystem
Since the attempts of controlling the servo motor used for releasing the messenger
based on the ultrasound range sensor measurements were not successful, it was decided
to use the flight controller for controlling the servo with pre-set commands. The Pixhawk
flight controller is capable of acquiring data from various sensors and controlling
different devices such as servo actuators (Gade et al., 2016). Servos are commonly used
as camera shutters or fixed winged model aircraft control (Polo et al., 2015). The Mission
Planner ground control software enables the servo control during an autonomous flight.
With the help of the flight controller, the servo can be triggered at any pre-determined
location at any desired altitude.
The UAV can fly autonomously using its autopilot function and automatically
obtain altitude information from its sensors (Chao et al., 2010); therefore, the water
sampler release action was controlled with autopilot commands. A “do-set-servo”
command was added to the flight mission at a desired location and altitude. This allowed
servo to be triggered at the sampling location then water sampler was released into
waterbody to capture water sample.
Water Sampler Performance Tests
The system components were tested indoor with the flight performance test
station; adjustments were made until the desired flight performance was achieved. The
water sampler was also tested outdoors to confirm the functions during an autonomous
29
mission. Outdoor experiments were conducted at a courtyard before using the system for
water sampling from a pond. A water tank with 0.6 m3 capacity was used to collect water
samples using the UAV. The depth of the water in the tank was 0.7 m. The surface area
of the water in the tank was 0.8 m2. Each UAV-assisted sampling attempt was recorded
based on their success. A total of 20 attempts were made to collect water samples with
the help of a remote controller. In order to determine the success rate of the sampler
mechanism, the UAV was remotely operated using the Loiter mode. The UAV was
launched from a location with a good GPS reception. Loiter mode allowed for relatively
easy flight control over the water tank. The UAV was launched and controlled with the
radio controller until the water sampler was inside the water tank. After the water sampler
dropped into the tank, approximately 0.4 m, the messenger was released manually by
turning a designated switch on the radio controller. The servo motor was rotated so that
the messenger would be released. The servo motor rotated between 0° and 40°. For the
initial tests, the servo motor was able to move either from its 0° position to 40° position
or from its 40° position to 0° position. Rotating from 40° to 0° position allowed a pin
holding the messenger to move backwards and release the messenger.
After releasing the messenger, the UAV ascended and returned to the launching
location. If the sampler captured water, it was recorded as a success along with the
amount of water. If the water sampler did not capture any water, it was recorded as a
failure and the reasons for the failure were recorded.
30
Experiment Location
The system was tested for its water sampling performance and autonomous flight
quality over Lamaster Pond at Clemson University, Clemson SC (Figure 6). Lamaster
Pond was within a 5 mile (8 km) radius of the Oconee County Regional Airport, SC. The
latitude and longitude of Lamaster Pond are 34.6578544º and 82.8193253º, and the
elevation of the pond is 709 ft (216 m) above the sea level according to the Clemson
USGS quad topographic map (Topozone, 2016). Because the UAV system did not have
enough endurance capacity to travel every corner of the pond, the middle of the pond was
selected for the outdoor experiments as the sampling location. The center of the pond also
allowed easy access by boat so that a traditional water sampling could be made to
confirm the sampling results from the UAV-assisted water sampling system within a
short time with relatively low cost. The launch location for autonomous missions was
represented with the letter “H” and the sampling location was represented with numbers
as shown in Figure 7. Numbers representing sampling location may vary due to the
number of commands selected for different purposes in the mission planner (Appendix
C). The distance between launch location and sampling point was 52m.
Figure 7. Location of the Lamaster Pond at Clemson University campus.
31
Regardless of their usage or purpose, all UAVs must be registered with the FAA
(Rainer, 2015). A UAV certificate number provided by FAA must be carried during
operations and placed on the UAV. The UAV certification number for the aircraft used
for this research is FA39H4CR3R. This number is valid until January 13, 2019.
Autonomous Flight Missions for Water Sampling
The schematic of water sampling operation is shown in Figure 8. At first, the
sampling mechanism was set, the flight controller and the power unit were connected, the
motors were armed and the UAV was set to idle mode. In this mode, the UAV was ready
to be launched by radio controller to start its mission. Next, mission mode was selected
from the radio controller, and following that the flight controller searched for a pre-set
mission. The throttle level was brought to 50% to provide enough thrust to the motors to
start the mission. As soon as the motors started spinning at 50% throttle, the UAV started
to ascend to the preset altitude and navigated to the predefined location. Once the UAV
reached to the sampling location, it descended to the sampling altitude (1.0 m) and
hovered for 18s at this altitude. During this time, the subsystem mounted on the UAV
released the messenger weight to trigger the water sampler. After 18 seconds, water was
captured and the UAV ascended to the navigation altitude to return to the launching
location to complete the water sampling mission.
32
Figure 8. Schematic showing the operation of the UAV-assisted water sampling.
Water samples were collected and on-site measurements were made at the pond.
Measured data was used to confirm that the UAV-mounted water sampling mechanism
did not bias the water quality indices. These indices were Dissolved Oxygen (DO),
temperature, electrical conductivity (EC), pH, and chloride. Temperature, pH, and DO
measurements need to be made in-situ or on-site because these constituents cannot be
analyzed adequately after transporting the sample to a laboratory (Paschke, 2003).
Therefore, these parameters were measured on-site with hand held measurement sensors.
Chloride ions were measured in the lab since these properties do not change rapidly after
33
sampling and filtering. Because DO is a key indicator of biological activity (Chung and
Yoo, 2015), it is possible that the sampling mechanism might bias the measurement
during sampling. Inland chloride ions naturally occur from dissolved minerals in most
surface waters, and they come from lawn fertilizers and road salt (Park et al., 2016).
Chloride ions are mobile in the soil and can be taken up by roots and accumulate in
leaves resulting in toxicity. High concentration of chloride in organisms can induce
osmotic stress and reduce fitness or mortality (Schulte, 1999). EC measurements are used
to identify existing problems in surface water for water quality monitoring. EC is an
indicator of the presence of mineral salt and fertilizers, organic matter, and treated
wastewater can contribute salt.
Measurement Procedures of Water Quality Indices
Water quality indices including pH, DO, EC, temperature, and Chloride were measured
from both handheld and UAV-assisted water samples. A Sension156 portable multi-
meter (Hach, USA) were used to measure pH and EC. A HQ10 portable DO meter (Hach,
USA) was used to measure DO and temperature. Chloride measurements were made
using a DR/2400 spectrophotometer (Hach, USA). Each collected water sample was
poured into a beaker for in-situ measurements and results were recorded manually.
Probes were rinsed with deionized water within each sampling and during calibration. A
field notebook of water sampling and on-site analysis information sheet was created
(Appendix D). Water samples were stored after in-situ measurements and transported to
the lab for chloride measurements. Water samples were labeled as “AS” for UAV-
34
assisted samples and “HS” for hand samples. A number following these labels indicated
the sample number (i.e. AS-26).
Calibration of pH and EC multi-meter and DO meter
Before water quality measurements were made, the instruments were calibrated in
order to receive accurate measurements. Two-point calibration was made for pH meter
using 4.01 pH and 7.00 pH buffers using the manufacturer’s manual. Daily calibration
accuracy checks were made by returning the electrode to a calibration buffer (4.01 pH) to
ensure that the pH meter provided accurate measurements.
Conductivity calibration was made with a known NaCl standard solution (1000
µS/cm at 25 °C). Calibration with this standard solution provides accuracy of the
measurements with conductivity of 0 to 10,000 µs (10 µS/cm) according to the
manufacturer’s manual. Conductivity ranges of common solution for deionized water was
1 µS/cm to 80 µS/cm. The DO probe did not require calibration by the operator for
manual operations. Therefore, the DO meter was used with its factory calibration.
Chloride measurement procedure
Mercuric thiocyanate method was used for Chloride measurements using the
manufacturer’s procedure. Samples were filtered using QT filters (Hach, USA) with 3-5
µm pore size and 125 mm diameter before analyses. Water samples were poured into 25
mL sample cells for analysis in DR/2400 spectrophotometer. These sample cells were
cleaned and wiped out before each analysis.
35
The chloride program was selected in the spectrophotometer menu. A sample cell
was filled with 25 mL of water sample. Another sample cell was filled with deionized
water to be used as a blank. Two mL of Mercuric Thiocyanate solution was added into
each sample cell. After sample cells were swirled to mix the solution properly, 1 mL of
Ferric Ion Solution was added into each sample cell and swirled to mix. Orange color
development was observed and a two minute of reaction time was allowed. Within 5
minutes of reaction time, blank was placed into the cell holder in the spectrophotometer
to zero the meter. Next, the prepared sample was wiped out and placed into the
spectrophotometer for Chloride measurement. This procedure was used to measure the
chloride in each sample and results were recorded manually.
36
4. RESULTS AND DISCUSSIONS
Payload Capacity of the UAV
The theoretical thrust of UAV can be calculated, so the determination for a
payload can be made. According to manufacturer’s specifications, MN3110 type 700KV
brushless DC motors (T-Motor, USA) can produce 430g of thrust with 10*45CF
propeller size at 50% throttle when 14.8 V battery was used. Total theoretical thrust that
the UAV can produce will be the sum of the thrust at each motors. Based on
manufacturer’s specifications the theoretical thrust that the UAV can produce was
calculated as 2,580g.
Since the external forces effects the basic motions of the UAV, it was decided to
measure actual thrust that the UAV can produce. Thrust measurement provided more
detailed information than the theoretical thrust calculation. It was essential to know the
actual payload capacity of the UAV. The change of thrust and endurance with throttle
(PWM duty cycle) were recorded during indoor testing (Figure 9). A 14.8V Li-Po battery
was used to estimate the average thrust that the hexacopter can produce at throttle levels
from 40 to 100% with total payload. The size of the battery used for the tests was 5000
mAh capacity with a 25C discharge rate.
37
Figure 9. UAV indoor flight performance test results.
The total weight of the UAV with the payload was 2730 g (Table 3). When the
throttle level increased, the produced thrust also increased while endurance decreased.
Indoor flight performance test results showed that the UAV was able to produce an
additional 1050.4 g of thrust at 50% throttle in lab conditions (Table 4). These test results
proved that the UAV could fly autonomously for 5.57 min when the throttle level for
autonomous flight was set at 50% (Figure 10). Therefore, the UAV must be able to
produce a minimum of 2730 g of thrust for lift off.
Table 3. Total weight of the hexacopter with payload.
Component Weight (g)
Hexacopter 2,333
Sampler 67
Messenger 200
Water sample 130
TOTAL 2730
0
1
2
3
4
5
6
7
8
300
500
700
900
1100
1300
1500
1700
1900
2100
2300
30 40 50 60 70 80 90 100 110
En
du
ran
ce (
s)
Th
rust
(g
)
Throttle (PWM Duty Cycle, %)
Thrust Endurance
38
Table 4. Average thrust and endurance at different throttle levels.
Throttle (%) Additional thrust (g) Endurance (s)
40 490.813 7.13
50 1050.40 5.79
60 1458.007 5.16
70 1719.217 4.37
80 1914.733 4.03
90 2026.23 3.67
100 2130.735 3.4
Figure 10. Change of thrust with endurance of the UAV at 50% throttle level.
Theoretical Endurance of the UAV
Theoretical flight time was calculated using Equation 7. The maximum capacity
of the battery was 5000 mAh and the amount of current pulled by the motors was 30 A.
The theoretical maximum flight time was calculated as 0.167 h (10 min). In order to
protect the battery life, only 80% of the battery needs to be used (Jeremia et al., 2012).
Therefore, the theoretical flight time was estimated to be 0.134 h (8 min).
500
600
700
800
900
1000
1100
1200
1300
1400
1500
1600
0 50 100 150 200 250 300 350 400
Thru
st (
g)
Endurance (s)
Test 1 Test 2 Test 3
39
Theoretical flight time was calculated using Equation 7:
pulled
fI
IT max (7)
Where,
Tf = Theoretical flight time (h)
Imax = Max current (A/h)
Ipulled = Pulled current from the battery (A)
Experimental Results of Water Sampler Performance
The average amount of water the sampler captured was 133.83 g (Table 5). Total
of 20 tests were conducted at the courtyard with a water tank. Twelve out 20 tests were
successful and 8 of them were failed. The major reasons for the water sampler not
capturing water during the water collection attempts were either due to the messenger not
being released or sampler not being triggered. The first one was because of the servo
malfunction due to tension caused by the weight of the messenger. The second reason for
failure was because of the velocity reduction of the messenger during the entry into
water. Messenger lost its energy during this entry and it did not apply enough force to
trigger the sampler. These results were used to estimate a probability of success for
autonomous outdoor experiments from sampling location. The probability of a successful
autonomous flight was estimated as 60%. In order to improve the probability of success
rate during autonomous water sampling, a servo actuator with a metal gear (Hextronic
HX12K with 1.977 N-m) was replaced with the plastic geared servo.
40
Table 5. Number of successful and failed sampling flights during the initial tests.
Result No of flights Explanation Avg. water captured (g)
Successful 12 Sampling depth was approximately 0.40m 133.83
Failed 8 Sampler malfunction 0
Estimation of Water Sampling Depth
Water sampling depth is important for the water quality measurements. Shallow
samples are generally collected from surface waters at a depth of 0 - 25 cm or from the
surface, and deep samples are taken from surface waters at a depth of more than 100 cm
(Wang, 2015). Before the UAV was sent for an autonomous sampling mission, the water
sampling altitude must be determined. Once the water sampling altitude was determined,
sampling depth could be estimated. The water sampler was connected to the messenger
release mechanism with a 1.8 m-long string. Observations from indoor and initial outdoor
experiments showed that the sampling depth must be between 0.3 and 1.0 m to allow the
messenger to function properly. The altitude measurements were made by the barometer
on the flight controller. The barometric pressure fluctuates with atmospheric conditions.
These fluctuations are higher during ascending and descending of the UAV (Zaliva and
Franchetti, 2014). Since additional payloads and external forces interfere with UAVs
attainable altitude range (Barve and Patel, 2014), the true altitude of the UAV must be
determined. Even though barometric pressure fluctuates with atmospheric conditions,
they are effective tools for determining the altitude of a moving vehicle and they are not
susceptible to problems by obstruction of satellite even in urban conditions (Zaliva and
Franchetti, 2014). Therefore, barometer altitude data from the control tuning (CTUN)
register was used to determine the altitude of the aircraft during water sampling events.
41
CTUN is a data registrar in Mission Planner contains throttle and altitude information of
each flight.
Barometric altitude data collection was made while flying the hexacopter in “auto
mode” during a water sampling mission (Figure 11). The data was downloaded from the
Pixhawk using Mission Planner and MAVLink protocol (QGroundControl, 2015). All of
the saved altitude variables were retrieved for identifying and selecting the ones that
represented the sampling altitude (Ardupilot, 2016). The control tuning (CTUN) register
which provides data for the throttle and the altitude was used to gather altitude data
(Salvador et al., 2015).
Figure 11. UAV with water sampler hovering at water sampling altitude during an autonomous flight
(left) and autonomous flight path for aerial water sampling (right).
A statistical analysis was conducted with SAS® to estimate the actual altitude of
the aircraft during sampling. Analysis on recorded CTUN barometer altitude was
conducted to determine if the average altitude was 1.0 m above the water surface during
18 s of water sampling duration. The total autonomous flight time was 1.5 min and only
18 s of this time was used for hovering over water surface during sample collection
(Figure 12). If the UAV hovers at 1.0 m altitude during water sampling, the water
42
sampling depth would be reported as 0.8 m. The data from three independent runs were
retrieved from Matlab® output. In each of the three independent runs, there were 180
altitude measurements. These measurements were not independent replicates. Data was
recorded with 100 μs intervals (Ardupilot, 2016). In this case, these 180 barometer
altitude data points represent the altitude measurements during the 18 s of hovering.
Weather conditions during the autonomous flight experiments are shown in Table 6.
Weather condition information was retrieved from a commonly available smartphone
application named “Weather” (Apple, USA).
Table 6. Autonomous flight information for altitude range experiments.
Data Log
Number
Date Time Temperature
(°C)
Wind Speed
(kmh)
Humidity
(%)
First 17-26-11 05.03.2016 4:47pm 21.67 22.53 44
Second 20-04-41 05.04.2016 5:11pm 27.78 11.27 57
Third 17-30-15 05.09.2016 5:33pm 27.78 14.48 33
Figure 12. Barometer altitude measurements during 18 seconds of water sampling from the first
autonomous flight.
A mixed procedure was followed to test if there were differences in the barometer
altitude in different flights using SAS® Studio (Appendix E). The null hypothesis was
43
that the aircraft’s altitude was 1.0 m during aerial water sampling at all times. The
alternative hypothesis was that the aircraft’s altitude was different than 1.0 m during
aerial water sampling at all times. Based upon SAS® output using the least square means
(Table 7), there was no sufficient evidence to conclude that the mean aircraft altitude was
equal to 1.0 m using 0.05 level of significance. The null hypothesis was rejected and the
alternative hypothesis was accepted which indicated that average aircraft altitude was
different than 1.0 m during water sampling, when it was set for 1.0 m as the sampling
altitude. Based on these results, the desired sampling depth was 66% satisfied. Water
sampling depth for UAV-assisted autonomous water sampling was between 0.79 m and
0.56 m based on outdoor experiments.
Table 7. Least Squares Means results from SAS®.
Least Squares Means
Flight No Estimate Standard Error DF t Value Pr > |t|
1 1.0213 0.01691 537 1.26 0.2076
2 1.2408 0.01691 537 14.24 <.0001
3 1.0121 0.01691 537 0.72 0.4741
Chloride Accuracy Check
Standard additions method was used to check the accuracy of chloride
measurements using the manufacturer’s procedure manual. Three sample spikes were
prepared using the chloride measurement procedure. Chloride standard solution of 0.1
mL, 0.2 mL, and 0.3 mL of 1000-mg/L was added to these sample spikes with the order
shown in Table 8. Each sample spikes were swirled to mix the solution and analyzed in
44
the same order. An accuracy graph was created using an ideal line option in the
spectrophotometer (Figure 13). Relationship between the sample spikes and the ideal line
showed 98.53% recovery.
Table 8. Chloride accuracy check and percent recovery.
Replicate Sample, mL Standard, mL CL, mg/L Recovery, %
Blank 0.25 0 5.7 100
1 0.25 0.1 10.5 108.9
2 0.25 0.2 12.9 95.3
3 0.25 0.3 17.5 100.7
Figure 13. Chloride accuracy chart.
Performance of Water Sampler during Autonomous Water Sampling
Missions for collecting water samples were recorded as successful missions, and
missions that did not collect water samples were recorded as unsuccessful missions. A
total of 73 autonomous water sampling flights were launched. Pre-flight checks for
autonomous water sampling missions were made before each flight. Total flight time, air
temperature, wind speed, and humidity was recorded for each day (Appendix F).
y = 37.8x + 5.98
R² = 0.9853
4
6
8
10
12
14
16
18
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
Ch
lori
de
(mg
/L)
Standard (mL)
Chloride (mg/L Cl)
45
48 of the 73 flights were recorded as successful missions. The probability of a successful
autonomous flight was estimated as 66%. Twenty of the unsuccessful water sampling
missions were due to messenger malfunction and 4 of the unsuccessful missions were due
to servo motor malfunction. Upgrading the servo motor from plastic gear to a metal gear
eliminated the malfunction issue, therefore no more servo malfunction issue was
observed. Only one of the missions were unsuccessful due to UAV stability. Wind speed
was greater than 28.97 kmh during this test. At this wind speed the stability of UAV was
weak and it crushed during landing. It is safe to recommend that the allowable maximum
wind speed for safe autonomous water sampling for this system is 24.14 kmh. However,
more research needs to be done to determine allowable wind speed limits for water
sampling system developed for this research.
In order to determine if any bias caused by sampler equipment, five different
water sampling locations were chosen in the pond (Figure 14). These five locations were
chosen as apart as possible from each other. Measurements were made only in the large
section of the pond to avoid accidents and for safe operation. The coordinates of each
sampling location can be found in the appendix (Appendix G).
46
Figure 14. Aerial image which was taken at 120 m altitude represents the sampling points.
Three handheld and three UAV-assisted water samples were collected at each
sampling point. DO, temperature, EC, and pH measurements were made on-site for both
methods since these values can change quickly during transportation. Chloride
measurements were made in the lab from 130 ml of samples which were collected with
both methods. The measured water quality indices, from hand-held samples and UAV-
assisted samples, from five different sample locations show the same general trend in all
of the five locations (Figure 15). This indicates that water sampling equipment and time
delay between sampling times had little impact on the water properties.
47
Figure 15. Water constituent measurements from hand sampling and UAV-assisted sampling. The points
represent the average of three replicate measurements for temperature (a), pH (b), DO (c), EC (d), and
Chloride (e).
The total time for unpacking the system and setting up the ground station, launching
missions for 15 water samples at five locations, analyzing after samples, and packing the
system was approximately one hour. Total time that took for accessing the five water
sampling location by kayak, obtaining water samples by hand, and analyzing them on-site
25.00
26.00
27.00
28.00
29.00
30.00
1 2 3 4 5
Tem
per
atu
re (°F
)
Location
a
UAV Hand
5.00
5.50
6.00
6.50
7.00
7.50
1 2 3 4 5
pH
(p
H)
Location
b
UAV Hand
9.20
9.40
9.60
9.80
10.00
10.20
10.40
1 2 3 4 5
DO
(m
g/L
)
Location
c
UAV Hand
67.00
68.00
69.00
70.00
71.00
72.00
73.00
74.00
1 2 3 4 5
EC
(µ
S/c
m)
Location
d
UAV Hand
3.50
4.00
4.50
5.00
5.50
6.00
6.50
1 2 3 4 5
Ch
lori
de
(mg
/L)
Location
e
UAV
Hand
48
was approximately 1 hour 10 minutes. These sampling points were close to each other. In
addition, the surface area of the pond was 11,000 m2 (Appendix G), which was not so
large. Therefore, sampling time was not an important parameter for determining the
quality of water sampling system in this case. A larger or not easy to access water body
can be chosen to determine if the UAV-assisted water sampling system reduces the
overall sampling time. In addition, this system can land on water surface and obtain water
sample while floating on the water. This landing option during water sampling might
provide advantages during water sampling at fast flowing waters but further experiments
are necessary.
Consistency between Manual and UAV-Assisted Sampling
In order to determine the consistency between manual and UAV-assisted
sampling, water samples from a location were collected for fourteen days. Water samples
were collected between May 9 and May 27. The first 9 samples and the last 5 samples
were collected continuously. Because of weather conditions, water samples were not
collected between May 18 and May 22. Three handheld and three UAV-assisted water
samples were collected at each sampling day. DO, temperature, EC, and pH
measurements were made on-site for both methods. The average of three measurements
represent water quality indices for each day. Chloride measurements were made in the
lab. Measurements from handheld samples and UAV-assisted samples are shown in
Figure 16. UAV-assisted water samplings were made at the sampling location 1 and
handheld water samplings were made near the shore, 3 meters away from sampling
49
location 2 (Figure 14). A paired t-test was used to compare the two population means for
each water quality constituent. The measured data for each method appeared to be
normally distributed. The null hypothesis was that the UAV-assisted water sampling
mechanism doesn’t induce bias in water quality measurements. The alternative
hypothesis was that the UAV-assisted water sampling mechanism induce bias in water
quality measurements. Based upon SAS® output (Appendix H), there was sufficient
evidence to conclude that the mean Chloride, temperature, and EC measurements made
with UAV-assisted water sampling method were similar to the handheld water sampling
method using 0.05 level of significance. The paired t-test results indicated that the UAV-
assisted water sampling mechanism does not induce bias in temperature, chloride and
electrical conductivity. However, the differences for pH and DO measurements for water
samples collected with UAV and handheld methods were significantly different at 0.05
level of significance. These differences might be due to time difference between two
sampling methods. There was at least one-hour time difference between two different
methods. Overall, temperature, EC, and Chloride measurements showed the same pattern.
DO and pH concentrations (Figure 16) revealed some differences between handheld
sampling and UAV-assisted sampling. These differences could be due to typical sampling
variations as well. Chloride values are usually expected to be in between 10 – 100 mg/L
(Ore et al., 2015). We can conclude that observed chloride variation was minimal.
50
20
22
24
26
28
30
32
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Tem
per
atu
re (
C)
Day
UAV Hand Air temp (C)
5
5.2
5.4
5.6
5.8
6
6.2
6.4
6.6
6.8
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
pH
(p
H)
Day
b
UAV Hand
8
9
10
11
12
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
DO
(m
g/L
)
Day
c
UAV Hand
51
Figure 16. Interaction plots between hand and UAV-assisted water constituent measurements for temperature (a), pH
(b), DO (c), EC (d), and chloride (e).
The experiments showed that more research needs to be conducted to determine
whether the UAV-assisted water sampling equipment can be replaced with handheld
water sampling method. This would provide more reliable data and further water quality
monitoring of Lamaster pond.
62636465666768697071727374
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
EC
(µ
S/c
m)
Day
d
UAV Hand
2
3
4
5
6
7
8
9
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Ch
lori
de
(mg
/L)
Day
e
UAV Hand
52
5. CONCLUSIONS
Water sampling is an important component of quality monitoring of surface
waters. Collection of water samples from difficult to access waterbodies and harsh
environments would be challenging with a boat. An autonomous water sampling system
containing a hexacopter and water capturing mechanism was developed and tested with
indoor and outdoor experiments. The system provided safe, rapid, and accurate water
sampling. The system performed autonomous water sampling tasks by collecting an
average of 130 ml water sample per mission. The system was equipped with floating
foams so that the water sampling would be achieved via a radio controller or
autonomously after the UAV is landed on the water surface. In order to prevent the UAV
from crashing during windy weather conditions, landing on water surfaces for water
sampling may also be beneficial. The system was verified for its consistency regarding
water constituent measurements. Water properties of the samples collected by the UAV-
assisted water sampler correlate well with the water properties of the samples collected
by traditional manual sampling. The UAV-assisted water sampler can be used by field
operators and scientists for remote water sampling to improve the resolution of water
sampling.
The UAV-assisted water sampling didn’t reduce the time required for water
sampling because of the size of the pond used for this study. Set-up process for the UAV-
assisted water sampling, safety and pre-flight checks, and battery replacement in between
sampling mission flights took most of the time. The system would be more useful if it is
used for large water surface sampling which may require large amount of water sample.
53
6. FUTURE WORK
The UAV-assisted water sampling system was capable of collecting a single
water sample at a flight mission. It may require field personal to launch more than one
mission flight to obtain water samples from a sampling location. This is a limiting factor
for time efficiency. In these situations, the UAV will use most of its power just to
navigate to the sampling location and back to the launch location. Water sampler can be
upgraded in a way that it can collect more than one water sample during one mission
flight. In this case, the UAV could navigate to the sampling location, obtain at least three
water samples, and return to the launch location.
Sampler triggering action was made by using a 200g of stainless steel messenger.
This messenger adds up additional weight to the system. Instead, triggering action can be
made with placing a much lighter servo motor on the top of the water sampler. This servo
could be waterproofed by the anti-corrosion spray so that it could operate underwater.
Replacing the messenger with servos would reduce the payload. This gain can be used to
add more water sampler to the current system.
Another UAV with bigger payload capacity would increase endurance. Because the
propeller thrust of the UAV will increase, more stable flight can be achieved. In this case,
allowable wind speed for safe flight would be higher. This would give an opportunity to
the UAV for autonomous water sampling at high wind speeds.
New sensor platforms for altitude approximation can be used to improve
navigational problems at low altitudes for autonomous water sampling. Because water
sampling is made while UAV is hovering over water surface, more accurate altitude
54
approximation using a more accurate sensor might be beneficial. Even though this system
can collect water samples after landing on waterbody, it is not fully safe for electronic
circuits. Fully waterproofed UAV would be more effective for water sampling.
55
APPENDICES
56
Appendix A. LabVIEW® Programs used for load cell calibration and load measurement.
Figure A-1. LabVIEW® program for load cell calibration.
Figure A-2. LabVIEW® program for load measurements with load cell.
57
Appendix B. Microcontroller programing code for automatic messenger release.
#define trigPin 12
#define echoPin 11
#include <Servo.h>
Servo myservo;
boolean minHeightReached = false;
void setup () {
Serial.begin (9600);
pinMode(trigPin, OUTPUT);
pinMode(echoPin, INPUT);
myservo.attach(7);
}
void loop () {
int duration, distance,pos=0,i;
digitalWrite(trigPin, LOW);
delayMicroseconds(2);
digitalWrite(trigPin, HIGH);
delayMicroseconds(10);
digitalWrite(trigPin, LOW);
duration = pulseIn(echoPin, HIGH);
distance = (duration/2) / 29.1;
//Serial.print(distance);
//Serial.println(" cm");
if(distance<200 && minHeightReached == false)
{
myservo.write(40);
}
if(distance<200 && minHeightReached == true)
{
myservo.write(20);
delay(4000);
myservo.write(0);
}
else if(distance>=200){
myservo.write(40);
minHeightReached = true;
}
delay(1000);
} Figure B-1. Microcontroller programing code for automatic messenger release.
58
Appendix C. Autonomous flight mission path and water sampling commands.
Figure C-1. Missin Planner flight plan for autonomous water sampling.
59
Appendix D. Field notebook of water sampling and on-site analysis.
Table D-1. Field notebook of water measurements details.
Field notebook of water sampling
Site: LaMaster Pond Description: Near LaMaster Dairy Farm
Station: Description:
Dates: 05.09.16 to 05.28.16 Time: Between 4pm and 5pm
Weather conditions: Partially Cloudy
Samples Collected: Standard Chemistry X Microbiology: N/A
Sampling depth: Between 40cm and 100cm
Problems encountered/adaptions made during sampling: Water sampler mechanism malfunction. A time delay (18
sec) adjustment was made in flight plan. That allowed more time for servo to complete its action in case if there is a
tension caused delay at servo movement. Sampling altitude adjustments were made. Final sampling altitude was 1m.
Therefore, sampling depth was approximately 80cm.
Sample preservation and storage:
130ml plastic containers. Chloride can be stored at room temperature for at least 28 days.
Sample transport:
Transportation with a truck. Approximate transportation distance is 3 miles from sampling location.
On-Site Analysis
Variable Method Equipment no Sampling codes Units
Temp DO meter LDO HQ10 AS: Aerial Sample and HS: Hand Sample °C
DO DO meter LDO HQ11 AS: Aerial Sample and HS: Hand Sample mg/L
pH pH meter LDO HQ12 AS: Aerial Sample and HS: Hand Sample pH
EC EC meter Sension156 AS: Aerial Sample and HS: Hand Sample µS/cm
Notes on on-site analysis:
Samples were used for Temperature, Dissolved Oxygen, pH, and EC measurements on-site. Three UAV-assisted
water samples and three hand-held water samples stored in 130ml containers for Chloride measurements in lab. UAV-
assisted water samples were collected from middle of the pond. Hand-held water samples were collected from a
location near shore with a plastic bottle. Coordinates for UAV-assisted water sampling: Lat"34.657048", Long"-
82.819796". Coordinates for Hand-held water sampling: Lat"34.6568001", Long"-82.8196648".
Name Signature Date
Collector: Cengiz Koparan 05.02.2016
60
Appendix E. SAS output for sampling depth analysis
Model Information
Data Set WORK.BARALT42
Dependent Variable BarAlt
Covariance Structure Variance Components
Estimation Method REML
Residual Variance Method Profile
Fixed Effects SE Method Model-Based
Degrees of Freedom Method Containment
Class Level Information
Class Levels Values
Flight 3 First Second Third
Obs 180 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89
90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122
123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152
153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180
Dimensions
Covariance Parameters 2
Columns in X 4
Columns in Z 540
Subjects 1
Max Obs per Subject 540
Number of Observations
Number of Observations Read 540
Number of Observations Used 540
Number of Observations Not Used 0
Iteration History
Iteration Evaluations -2 Res Log Like Criterion
0 1 -53.80061646
1 1 -53.80061646 0.00000000
Convergence criteria met.
61
Covariance Parameter Estimates
Cov Parm Estimate
Obs (Flight) 0.002518
Residual 0.04894
Fit Statistics
-2 Res Log Likelihood -53.8
AIC (Smaller is Better) -49.8
AICC (Smaller is Better) -49.8
BIC (Smaller is Better) -41.2
Type 3 Tests of Fixed Effects
Effect
Num
DF
Den
DF F Value Pr > F
Flight 2 537 58.61 <.0001
Least Squares Means
Effect Flight Estimate
Standard
Error DF t Value Pr > |t|
Flight First 0.02133 0.01691 537 1.26 0.2076
Flight Second 0.2408 0.01691 537 14.24 <.0001
Flight Third 0.01211 0.01691 537 0.72 0.4741
62
Appendix F. Pre-flight check list before autonomous water sampling.
Table F-1. Pre-flight and post flight check list.
Pre-Flight Check List
May 2016 2 9 10 11 12 13 14 15 16 17 18 23 24 25 26 27
Visual check of airframe √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Motors free √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Battery secure √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
No loose leads √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
All antennas secure √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Propellers – no damage, secure √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Camera – secure √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Control RC – check battery/antenna √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Ground station powered and online √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
RC switch positions – throttle closed √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Battery voltage sufficient √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Power flight controller √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Camera operational – set √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Flight controller functional – GPS lock √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Lift off, check controls √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Altitude hold and loiter modes functional √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Auto mode started √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Post-Flight Check List
Observe normal motor rundown √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Record time √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Motor and ESC temperature normal √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Battery temperature normal √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Disconnect ground station √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Unplug the battery √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Turn-off RC √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Turn-off camera √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Visual check of airframe √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Motors free √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Battery secure √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
No loose leads √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
All antennas secure √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Propellers – no damage, secure √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Camera – secure √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
63
Table F-2. UAV flying log book for water sampling flights.
Log Book and Aircraft Information
FAA Registration Number: FA39H4CR3R
Aircraft Type: UAV-Hexacopter
Control System: Pixhawk
Radio Controller: Turnigy 9x
Wight: 3000g
Motors: T-Motor 700kv Brushless Dc
Propellers: 1045 Carbon Fiber
Esc: 30a 6cell
Battery: 14.4v 25c Discharge, 6cell Li-Po
Radio Communication: 915mhz Radio Telemetry
Endurance: ~6 Max With Payload. Date Time Operator Duration
(min-total)
Task Location Temp
(°F)
Wind
(mph)
Humidity
(%)
05.02.2016 4:40pm Cengiz 36 WS Lamaster 82 5-8 55-57
05.03.2016 4.16pm Cengiz 36 WS Lamaster 82 14-16 41-45
05.09.2016 4:50pm Cengiz 36 WS Lamaster 84 6-9 33
05.10.2016 4:19pm Cengiz 36 WS Lamaster 86 5-8 33-38
05.11.2016 4:56pm Cengiz 42 WS Lamaster 79 4-7 54-60
05.12.2016 4:50pm Cengiz 18 WS Lamaster 79 6-8 58-59
05.13.2016 4:10pm Cengiz 24 WS Lamaster 82 10-15 32-36
05.14.2016 4:10pm Cengiz 6 WS Lamaster 82 15-17 33-34
05.15.2016 4:25pm Cengiz 18 WS Lamaster 75 6-9 21-22
05.16.2016 4:10pm Cengiz 30 WS Lamaster 73 0-3 32-36
05.17.2016 3:54pm Cengiz 42 WS Lamaster 72 5-6 66-74
05.23.2016 4:30pm Cengiz 30 WS Lamaster 81 2-6 30-34
05.24.2016 4:26pm Cengiz 18 WS Lamaster 82 5-7 34-36
05.25.2016 5:04pm Cengiz 18 WS Lamaster 86 6-8 34-35
05.26.2016 4:10pm Cengiz 84 WS Lamaster 81 7-9 47-50
05.27.2016 4:10pm Cengiz 49 WS Lamaster 82 6-7 45-47
64
Appendix G. Autonomous water sampling locations and boundaries of sampling site.
Table G-1. Coordinates of water sampling locations.
Location Latitude (°) Longitude (°)
1 34.657086 -82.819755
2 34.656881 -82.819359
3 34.657249 -82.819702
4 34.657187 -82.819279
5 34.657388 -82.819472
Figure G-1. Boundaries of sampling site on the pond.
65
Appendix H. SAS Output for water sampling consistency analysis.
Table H-1. Chloride measurement consistency analysis
N Mean Std Dev Std Err Minimum Maximum
14 -1.0167 0.9393 0.2511 -2.5000 1.0667
Mean 95% CL Mean Std Dev 95% CL Std Dev
-1.0167 -1.4613 Infty 0.9393 0.6810 1.5133
DF t Value Pr > t
13 -4.05 0.9993
With 95% Upper Confidence Interval for Mean
Distribution of Difference: UAV - Hand
95% Confidence
0
10
20
30
40
Per
cent
0
10
20
30
40
Per
cent
Kernel
Normal
-4 -2 0 2
Difference
66
UAV Hand
8
9
10
11
8
9
10
11
Paired Profiles for (UAV, Hand)
UAV Hand
8
9
10
11
8
9
10
11
Mean
Paired Profiles for (UAV, Hand)
8 9 10 11
UAV
8
9
10
11
Han
d
Agreement of Hand and UAV
8 9 10 11
UAV
8
9
10
11
Han
d
Mean
Agreement of Hand and UAV
67
Appendix H-2. DO measurement consistency analysis
N Mean Std Dev Std Err Minimum Maximum
14 1.0167 0.9393 0.2511 -1.0667 2.5000
Mean 95% CL Mean Std Dev 95% CL Std Dev
1.0167 0.5721 Infty 0.9393 0.6810 1.5133
DF t Value Pr > t
13 4.05 0.0007
-2 -1 0 1 2
Quantile
-3
-2
-1
0
1
Diffe
renc
e
Q-Q Plot of Difference: UAV - Hand
68
With 95% Upper Confidence Interval for Mean
Distribution of Difference: UAV - Hand
95% Confidence
0
10
20
30
40
Per
cent
0
10
20
30
40
Per
cent
Kernel
Normal
-2 0 2
Difference
UAV Hand
8
9
10
11
8
9
10
11
Paired Profiles for (UAV, Hand)
UAV Hand
8
9
10
11
8
9
10
11
Mean
Paired Profiles for (UAV, Hand)
69
8 9 10 11
UAV
8
9
10
11
Han
d
Agreement of Hand and UAV
8 9 10 11
UAV
8
9
10
11
Han
d
Mean
Agreement of Hand and UAV
-2 -1 0 1 2
Quantile
-1
0
1
2
3
Diffe
renc
e
Q-Q Plot of Difference: UAV - Hand
70
Appendix H-3. EC measurement consistency analysis.
N Mean Std Dev Std Err Minimum Maximum
14 0.5810 2.1055 0.5627 -3.5000 4.4667
Mean 95% CL Mean Std Dev 95% CL Std Dev
0.5810 -0.4156 Infty 2.1055 1.5264 3.3921
DF t Value Pr > t
13 1.03 0.1604
With 95% Upper Confidence Interval for Mean
Distribution of Difference: UAV - Hand
95% Confidence
0
10
20
30
Per
cent
0
10
20
30
Per
cent
Kernel
Normal
-5.0 -2.5 0.0 2.5 5.0
Difference
71
UAV Hand
64
66
68
70
72
64
66
68
70
72
Paired Profiles for (UAV, Hand)
UAV Hand
64
66
68
70
72
64
66
68
70
72
Mean
Paired Profiles for (UAV, Hand)
64 66 68 70 72
UAV
64
66
68
70
72
Han
d
Agreement of Hand and UAV
64 66 68 70 72
UAV
64
66
68
70
72
Han
d
Mean
Agreement of Hand and UAV
72
Appendix H-4. pH measurement consistency analysis.
N Mean Std Dev Std Err Minimum Maximum
14 0.1131 0.3784 0.1011 -0.6333 0.8000
Mean 95% CL Mean Std Dev 95% CL Std Dev
0.1131 -0.0660 Infty 0.3784 0.2743 0.6096
DF t Value Pr > t
13 1.12 0.1418
-2 -1 0 1 2
Quantile
-4
-2
0
2
4
Diffe
renc
e
Q-Q Plot of Difference: UAV - Hand
73
With 95% Upper Confidence Interval for Mean
Distribution of Difference: UAV - Hand
95% Confidence
0
10
20
30
Per
cent
0
10
20
30
Per
cent
Kernel
Normal
-1.0 -0.5 0.0 0.5 1.0
Difference
UAV Hand
5.5
6.0
6.5
5.5
6.0
6.5
Paired Profiles for (UAV, Hand)
UAV Hand
5.5
6.0
6.5
5.5
6.0
6.5
Mean
Paired Profiles for (UAV, Hand)
74
5.5 6.0 6.5
UAV
5.5
6.0
6.5
Han
d
Agreement of Hand and UAV
5.5 6.0 6.5
UAV
5.5
6.0
6.5
Han
d
Mean
Agreement of Hand and UAV
-2 -1 0 1 2
Quantile
-0.5
0.0
0.5
1.0
Diffe
renc
e
Q-Q Plot of Difference: UAV - Hand
75
Appendix H-5. Temperature measurement consistency analysis.
N Mean Std Dev Std Err Minimum Maximum
14 1.3321 1.0872 0.2906 -0.1334 3.2000
Mean 95% CL Mean Std Dev 95% CL Std Dev
1.3321 0.8176 Infty 1.0872 0.7881 1.7515
DF t Value Pr > t
13 4.58 0.0003
With 95% Upper Confidence Interval for Mean
Distribution of Difference: UAV - Hand
95% Confidence
0
5
10
15
20
25
30
Per
cent
0
5
10
15
20
25
30
Per
cent
Kernel
Normal
-2 0 2 4
Difference
76
UAV Hand
24
26
28
30
24
26
28
30
Paired Profiles for (UAV, Hand)
UAV Hand
24
26
28
30
24
26
28
30Mean
Paired Profiles for (UAV, Hand)
24 26 28 30
UAV
24
26
28
30
Han
d
Agreement of Hand and UAV
24 26 28 30
UAV
24
26
28
30
Han
d
Mean
Agreement of Hand and UAV
77
-2 -1 0 1 2
Quantile
-1
0
1
2
3
Diffe
renc
e
Q-Q Plot of Difference: UAV - Hand
78
REFERENCES
3D Printing. 2016. Types of 3D printers or 3D printing technologies overview. Available
at: http://3dprintingfromscratch.com/common/types-of-3d-printers-or-3d-printing-
technologies-overview/. Accessed 21 June 2016.
Addy, K., L. Green, E. Herron and K. Stepenuck. 2010. Fact Sheet: Why Volunteer
Water Quality Monitoring Makes Sense. Available
at: http://acwi.gov/monitoring/vm/publications/articles/article_Why_Makes_Sense.pd
f. Accessed 29 November 2015. Advisory Committee on Water Information.
Ambati, P. R. and R. Padhi. 2016. Robust auto-landing of fixed-wing UAVs using neuro-
adaptive design. Control Engineering Practice.
Ardupilot. 2016. Downloading and Analyzing Data Logs in Mission Planner. Available
at: http://ardupilot.org/copter/docs/common-downloading-and-analyzing-data-logs-in-
mission-planner.html. Accessed 20 April 2016.
Atlam, O. and M. Kolhe. 2013. Performance evaluation of directly photovoltaic powered
DC PM (direct current permanent magnet) motor – propeller thrust system. Energy
57692-698.
Baránek, R. and F. Šolc. 2012. Modelling and control of a hexa-copter. IEEE Robotics &
Automation Magazine.
Bartram, J. and R. Ballance, eds. 1996. Water Quality Monitoring - A Practical Guide to
the Design and Implementation of Freshwater Quality Studies and Monitoring
Programmes. Chapter 5, Field Work and Sampling. World Health Organization.
Barve, J. and K. Patel. 2014. Modelling, Simulation and Altitude-Range-Analysis of
Quad-copter UAV. IFAC Proceedings Volumes 47(1): 1126-1130.
Campbell, B. A. and S. W. McCandless Jr. 1996. Chapter 6 - Remote Sensing. In
Introduction to Space Sciences and Spacecraft Applications, 132-153. ed. B. A.
Campbell, and McCandless, S. W., Houston: Gulf Professional Publishing.
Chao, H., Y. Luo, L. Di and Y. Q. Chen. 2010. Roll-channel fractional order controller
design for a small fixed-wing unmanned aerial vehicle. Control Engineering Practice
18(7): 761-772.
Chung, W. Y. and J. H. Yoo. 2015. Remote water quality monitoring in wide area.
Sensors and Actuators B: Chemical 21751-57.
79
Clayton Technologies, ed. 2013. HC-SR04 Ultrasound sensor user's manual. Available
at: https://docs.google.com/document/d/1Y-yZnNhMYy7rwhAgyL_pfa39RsB-
x2qR4vP8saG73rE/edit. Accessed on 12 February 2016. 1st ed.
Copeland, C. 2012. Clean Water Act and Pollutant Total Maximum Daily Loads
(TMDLs). Available at: https://www.fas.org/sgp/crs/misc/R42752.pdf. Accessed on 8
November 2016. R42752.
Cutler, M., N. K. Ure, B. Michini and J. P. How. 2011. Comparison of fixed and variable
pitch actuators for agile quadrotors. In .
Dlamini, S., I. Nhapi, W. Gumindoga, T. Nhiwatiwa and T. Dube. 2016. Assessing the
feasibility of integrating remote sensing and in-situ measurements in monitoring
water quality status of Lake Chivero, Zimbabwe. Physics and Chemistry of the Earth,
Parts A/B/C.
Doshi, A. A., A. J. Postula, A. Fletcher and S. P. N. Singh. 2015. Development of micro-
UAV with integrated motion planning for open-cut mining surveillance.
Microprocessors and Microsystems 39(8): 829-835.
Erkkilä, A. and R. Kalliola. 2004. Patterns and dynamics of coastal waters in multi-
temporal satellite images: support to water quality monitoring in the Archipelago Sea,
Finland. Estuarine, Coastal and Shelf Science 60(2)(165-177): .
FAA. 2016. Unmanned Aircraft Systems (UAS) Regulationsa & Policies. Available
at: https://www.faa.gov/uas/regulations_policies/. Accessed 20 February 2016.
Fortuna, J., F. Ferreira, R. Gomes, S. Ferreira and J. Sousa. 2013. 2nd IFAC Workshop
on Research, Education and Development of Unmanned Aerial Systems Using low
cost open source UAVs for marine wild life monitoring - Field Report*. IFAC
Proceedings Volumes 46(30): 291-295.
Gade, M. M., S. Hangal, D. Krishnan and H. Arya. 2016. Development of Obstacle
Avoidance Controller for MAVs: Testing in Hardware-In-Loop Simulation∗. IFAC-
PapersOnLine 49(1): 413-418.
Gupta, A. K. and V. Jha. 2014. Design and Development of Remote Controlled
Autonomous Synchronic Hexaroter Aerial (ASHA) Robot. Procedia Technology
14(0)(51-58): .
Herring, P. J., A. K. Campbell, M. Whitfield and L. Maddock. 1990. Light and Life in the
Sea. Cambridge: Cambridge Univ. Press.
80
Jeremia, S., E. Kuantama and J. Pangaribuan. 2012. Design and construction of remote-
controlled quad-copter based on STC12C5624AD. System Engineering and
Technology (ICSET).
Kaizu, Y., M. Iio, H. Yamada and N. Noguchi. 2011. Development of unmanned airboat
for water-quality mapping. Biosystems Engineering 109(4): 338-347.
Karimanzira, D., M. Jacobi, T. Pfuetzenreuter, T. Rauschenbach, M. Eichhorn, R.
Taubert and C. Ament. 2014. First testing of an AUV mission planning and guidance
system for water quality monitoring and fish behavior observation in net cage fish
farming. Information Processing in Agriculture 1(2): 131-140.
Lim, H. S. and X. X. Lu. 2016. Sustainable urban stormwater management in the tropics:
An evaluation of Singapore’s ABC Waters Program. Journal of Hydrology 538842-
862.
Liu, R., F. Xu, P. Zhang, W. Yu and C. Men. 2016. Identifying non-point source critical
source areas based on multi-factors at a basin scale with SWAT. Journal of
Hydrology 533(379-388): .
Lo, C. H., H. Shin, A. Tsourdos and S. Kim. 2013. Modeling and simulation of fault
tolerant strategies for a quad rotor UAV. In AIAA Modeling and Simulation
Technologies Conference. American Institute of Aeronautics and Astronautics, .
Luhtala, H. and H. Tolvanen. 2016. Spatio-temporal representativeness of euphotic depth
in situ sampling in transitional coastal waters. Journal of Sea Research 11232-40.
Ma, Y., P. Egodawatta, J. McGree, A. Liu and A. Goonetilleke. 2016. Human health risk
assessment of heavy metals in urban stormwater. Science of The Total Environment
557–558764-772.
Madrid, Y. and Z. P. Zayas. 2007. Water sampling: Traditional methods and new
approaches in water sampling strategy. TrAC Trends in Analytical Chemistry 26(4):
293-299.
Mahony, R., V. Kumar and P. Corke. 2012. Multirotor Aerial Vehicles: Modeling,
Estimation, and Control of Quadrotor. IEEE Robotics & Automation Magazine 19(3):
20-32.
Massa, D. P., ed. 1999. Choosing Ultrasonic Sensor for Proximity or Distance
Measurment. Massa Products Corp.: Helmers Publishing, inc.
McGowan, S., J. F. Sivanpillai and S. Ramesh. 2016. Algal Blooms In Biological and
Environmental Hazards, Risks, and Disasters, 5-43. Boston Academic Press.
81
Myers, M. D. 2006. Collection of water samples. Available at:
http://pubs.water.usgs.gov/twri9A/. Accessed 12 October 2015.: U.S. Geological
Survey.
Nation, T. H. and L. A. Johnson. 2015. Use of a Volunteer Monitoring Program to Assess
Water Quality in a TMDL Watershed Utilized for Recreational Use, Pickens County,
South Carolina. South Carolina Water Resources 2(1): 11-15.
National Instrument. 2013. What is LabVIEW?. Available
at: http://www.ni.com/newsletter/51141/en/. Accessed 28 May 2016.
Neumann, M., M. Liess and R. Schulz. 2003. A qualitative sampling method for
monitoring water quality in temporary channels or point sources and its application to
pesticide contamination. Chemosphere 51(6)(509-513): .
Nilssen, I., Ø Ødegård, A. J. Sørensen, G. Johnsen, M. A. Moline and J. Berge. 2015.
Integrated environmental mapping and monitoring, a methodological approach to
optimise knowledge gathering and sampling strategy. Marine pollution bulletin 96(1–
2): 374-383.
Nishar, A., S. Richards, D. Breen, J. Robertson and B. Breen. 2016. Thermal infrared
imaging of geothermal environments and by an unmanned aerial vehicle (UAV): A
case study of the Wairakei – Tauhara geothermal field, Taupo, New Zealand.
Renewable Energy 861256-1264.
Ore, J. P., S. Elbaum, A. Burgin, B. Zhao and C. Detweiler. 2015. Autonomous Aerial
Water Sampling. Available at: http://cse.unl.edu/~carrick/papers/OreEBZD-
FSR2013.pdf. Accessed 20 December 2015.
Park, D. M., L. B. McCarty and S. A. White. 2016. Interpreting Irrigation Water Quality
Reports. Clemson University Water Chemistry.
Paschke, A. 2003. Consideration of the physicochemical properties of sample matrices–
an important step in sampling and sample preparation. TrAC Trends in Analytical
Chemistry 22(2): 78-89.
Patterson, T., S. McClean, P. Morrow, G. Parr and C. Luo. 2014. Timely autonomous
identification of UAV safe landing zones. Image and Vision Computing 32(9): 568-
578.
Polo, J., G. Hornero, C. Duijneveld, A. García and O. Casas. 2015. Design of a low-cost
Wireless Sensor Network with UAV mobile node for agricultural applications.
Computers and Electronics in Agriculture 11919-32.
82
QGroundControl. 2015. MAVLink Micro Air Vehicle Communication Protocol.
Available at: http://qgroundcontrol.org/mavlink/start. Accessed 15 April 2016.
Rainer, D. 2015. Rules, Regulations and Codes for Drones, Unmanned Aerial Vehicle,
NextGen Air Transportation, Unmanned Air Systems. Journal of Chemical Health
and Safety 22(3): 37-39.
Rockwell Automation. 2016. Ultrasonic Sensing. Available at:
http://www.ab.com/en/epub/catalogs/12772/6543185/12041221/12041229/print.html.
Accessed February 14 2016.
Salvador, B., J. C. Moreno and J. L. Guzman. 2015. Modelling of a non-commercial
UAV for control and robotics laboratory. IFAC-PapersOnLine 48(29): 65-69.
SCDHEC. 2015. South Carolina Department of Environment ControI. Introduction page.
Available at: http://www.scdhec.gov/agency/. Accessed on 23 November 2015.
Schulte, E. E. 1999. Soil and applied chlorine. Understanding Plant Nutrients. Available
at: http://corn.agronomy.wisc.edu/Management/pdfs/a3556.pdf. Accessed 26 April
2016.
SCWRRI. 2004. Annual Technical Report. Available
at: http://water.usgs.gov/wrri/AnnualReports/2004/FY2004_SC_Annual_Report.pdf.
Accessed 7 September 2015.
Segui-Gasco, P., Y. Al-Rihani, H. Shin and A. Savvaris. 2014. A Novel Actuation
Concept for a Multi Rotor UAV. Available
at: http://link.springer.com/article/10.1007%2Fs10846-013-9987-3. Accessed 6 May
2016. Journal of Intelligent & Robotic Systems 74(1): 173-191.
Shan, J. 2016. Payload Design [Lecture Notes]. Available
at: http://www.yorku.ca/jjshan/eng4360/notes_files/ENG4360_2_1_Payload_Design_
Sizing.pdf. Accessed 22 June 2016.
Thomas, K. V., M. R. Hurst, P. Matthiessen, D. Sheahan and R. J. Williams. 2001.
Toxicity characterisation of organic contaminants in stormwaters from an agricultural
headwater stream in South East England. Water research 35(10): 2411-2416.
Topozone. 2016. Lamaster Pond Topo Map in Pickens County SC. Available
at: http://www.topozone.com/south-carolina/pickens-sc/reservoir/lamaster-pond/.
Accessed 15 April 2016.
Turner, I. L., M. D. Harley and C. D. Drummond. 2016. UAVs for coastal surveying.
Coastal Engineering 11419-24.
83
van der Merwe, D. 2015. Chapter 31 - Cyanobacterial (Blue-Green Algae) Toxins. In
Handbook of Toxicology of Chemical Warfare Agents (Second Edition), 421-429. ed.
R. C. Gupta, Boston: Academic Press.
VPG Transducers. 2015. Load Cell Technology. Available
at: http://www.vishaypg.com/docs/11866/vpg-01.pdf. Accessed 28 May 2016.
Wang, X. 2015. Reprint of “China geochemical baselines: Sampling methodology”.
Journal of Geochemical Exploration 15417-31.
Warren, N., I. J. Allan, J. E. Carter, W. A. House and A. Parker. 2003. Pesticides and
other micro-organic contaminants in freshwater sedimentary environments—a
review. Applied Geochemistry 18(2): 159-194.
Wilde, F. D., D. B. Radtke, J. Gibs and R. T. Iwatsubo, eds. 2003. National Field Manual
for the Collection of Water-Quality Data. Book 9, Chapter A2, Page 15. USGS-TWRI.
Available at:
https://water.usgs.gov/owq/FieldManual/Chapter2/Chapter2_V2uncompressed.pdf.
Accessed 20 May 2016.
Wilson, D. 2015. Integrated Report for 2014. Part 1: Seciton 303(d) List of Impaired
Waters. Available at: https://www.scdhec.gov/HomeAndEnvironment/Docs/tmdl_14-
303d.pdf. Accessed 10 October 2016.
Xiong, J. and E. Zheng. 2015. Optimal Kalman Filter for state estimation of a quadrotor
UAV. Optik - International Journal for Light and Electron Optics 126(21): 2862-
2868.
Zaliva, V. and F. Franchetti. 2014. Barometric and GPS Altitude Sensor Fusion. IEEE
Robotics & Automation Magazine.