AEROTHON 2024
Uncrewed Aircraft
System (UAS) Design, Build
And Fly Contest
Design Report
Team Classified
AT2024 - 013
Madras Institute of Technology, Anna
University, Chennai, Tamil Nadu - 600044
1
AeroTHON 2024 Team Classified
2
AeroTHON 2024 Team Classified
Table of Contents
1 Technical Content 4-21
1.1 Conceptual Design 4
1.2 Preliminary Weight Estimation 8
1.3 Preliminary Propulsion Estimation 8-9
1.3.1 Thrust Required Estimation 8
1.3.2 Propulsion System Selection 9
1.4 Aircraft Sizing 10-12
1.4.1 Rotor arm 10
1.4.2 Hub 10
1.4.3 Wheelbase 11
1.4.4 Propeller Clearance 12
1.4.5 Landing Gear 12
1.5 Aircraft Performance 13-14
1.5.1 Power Required Estimation 13
1.5.2 Power System Battery Selection 13
1.5.3 Endurance Estimation 14
1.6 Material Selection 15
1.7 Subsystem Selection 15-18
1.7.1 Communication System 15
1.7.2 Control and Navigation System 17
1.8 CG and Stability Analysis 19-20
1.8.1 CG Analysis 19
1.8.2 Stability Analysis 20
1.9 Preliminary Computer Aided Design Model 21
2 Computational Analysis 21-24
2.1 Optimized Final Design 25
2.2 Recalculation 26-29
2.2.1 Detailed Weight Breakdown 26
2.2.2 Aircraft Performance Recalculation 27
2.2.3 Final UAV Specifications 28
2.2.4 Bill of materials 29
3 Methodology for Autonomous Operation 30-31
3.1 Autonomous Flight 30
3.2 Autonomous Object Detection and Counting 30
3.3 Autonomous Payload Drop 31
4 Innovation 32-33
References 33
Appendix 34-40
3
AeroTHON 2024 Team Classified
1. TECHNICAL CONTENT
1.1 Conceptual Design
Quadcopters are helicopters equipped with four rotors, each contributing to the aircraft's flight
by generating upward thrust. Unlike traditional helicopters, all rotors on a quadcopter share the
responsibility of lift.
The thrust and torque are the very basic two things used for movement of quadcopter. The rotors
are arranged in pairs that rotate in opposite directions, ensuring that the net torque on the aircraft
remains zero. This configuration allows the quadcopter to maintain stability and control. With six
degrees of freedom (x, y, z, θ, ɸ, ψ), a quadcopter can move forward, backward, left, right, up,
and down, and it can also rotate around its three axes.
They typically have 2 configurations: X Frame and + Frame
We are choosing the True X Frame configuration for our drone because of its widely used
application for delivery drones
Mission Objective:
The objective is to design, build and fly a multirotor UAV that can deliver cargo to a specified
location along with survey, object identification and counting. The UAS has to carry a 200g
payload and deliver it to a target while performing survey and object identification with
dimensions of 10x5x5 cm as shown
4
AeroTHON 2024 Team Classified
High Level Component Overview
SIDE VIEW
GPS
Battery
Propeller
BLDC Motor
Transceiver
ESC
Camera
Pixhawk
5
AeroTHON 2024 Team Classified
FRONT VIEW
Telemetry transmitter
PDB
FPV Camera
Raspberry Pi Servos
Payload
Battery 1S
Working of Payload Dropping Mechanism
Base Plate
Servo Motors
Holders
Claws
Payload
6
AeroTHON 2024 Team Classified
Stowed away configuration for
payload in the payload dropping
mechanism
Release of payload after reaching
target at 5m altitude
Payload is dropped after the servo
motors rotates 35 degrees with
enough room to fall without
disturbance
7
AeroTHON 2024 Team Classified
1.2 Preliminary Weight Estimation
Here we have the preliminary weight estimated for all components
Component Qty Total Weight
BLDC Motor 4 190
ESC 4 104
Propellers 4 20
GPS module 1 26
Power Distribution Board 1 11
Companion Controller 1 52
Flight Controller 1 40
Telemetry Transmitter 1 33
Battery 1 550
Camera, Servo, Wires - 125
Payload 1 200
Custom built parts - 560
(Mechanism + Frame)
Our preliminary total weight comes out to be 1906g. We will use this moving forward with the
calculations
1.3 Preliminary Propulsion Calculations
1.3.1 Thrust required estimation
Our preliminary gross weight for the drone has been assumed to be around 1906g. We have to
pick our thrust-to-weight ratio to find the optimal thrust required to fly the drone. The higher the
thrust-to-weight ratio, the easier it is to control your drone in elaborate aerobatics. A 2:1 thrust-
to-weight ratio will allow our drone to hover at just half throttle.
Assumed weight = 1906g
Which is the Minimum Thrust required to hover.
Total weight in Newtons = 1.906 * 9.81 = 18.69786 N
Total Thrust required with assumed 2:1 ratio = 37.39572 N
Thrust per Propeller = 37.39572/ 4 = 9.34893 N
So, each motor-propeller combination should produce a minimum thrust of 9.34 N or 0.953 Kgf.
8
AeroTHON 2024 Team Classified
1.3.2 Propulsion system selection
For the propulsion system, we have chosen an electrically driven propeller system. This
system uses a combination of four brushless DC motors and propellers to achieve the
requirement of 37.39 N or 3.812 Kgf of thrust.
A process diagram for the working of the propulsion system is given below:
Motors –
The motors that we have selected, the Emax ECO II Series
2807 1500KV is an excellent choice for this application due
to is efficiency and reliability.
It has a maximum RPM of 22200 as seen from the datasheet
for 4S configuration.
Electronic speed controller –
Electronic Speed Controllers or ESCs are used to
control and adjust the speed of our motors. The motor
we have chosen is a 1500kv motor and has a maximum
current draw of 27.65 A.
This thus necessitates the use of a 40 Amp ESC
ReadytoSky 40A 2-6S ESC is chosen for meeting the
above requirements and being compatible with the
motors and battery.
9
AeroTHON 2024 Team Classified
Propellers –
For the propellers we have chosen the Orange HD 8045
propeller which is an 8-inch 4.5-inch pitch Carbon Nylon
propeller.
It was chosen due to is lightweight and durable
construction as it is constructed out of Carbon Nylon,
and it is readily compatible with the Emax ECO II.
1.4 Aircraft Sizing
1.4.1 Rotor Arm
The rotor arm of a drone in an X configuration is a crucial component that contributes to the
overall stability and control of the drone. The design of the rotor arm with dimensions is given
below, which is enough for the arm to withstand thrust generated by the motor and force of
impact from the landing gear.
1.4.2 Hub
The hub in a drone is a central component that is responsible in maintaining structural integrity,
and housing all the electronic components. The dimensions of the Hub in this case refers to the
top and the base plate whose dimensions are given below. The base plate has attachments that
also belong to the payload dropping mechanism.
10
AeroTHON 2024 Team Classified
Base Plate
Top Plate
1.4.3 Wheelbase
Wheelbase refers to the distance between the centers of the motors on opposite sides of the
drone. It's a measurement of the span of the drone's arms, indicating how far apart the motors
are from each other. For our drone’s design, the wheelbase dimension is 419.22 mm.
11
AeroTHON 2024 Team Classified
1.4.4 Propeller clearance
Propeller clearance is the space between the tips of the propellers and other parts of the drone,
such as the drone's frame, body, or other propellers.
For our quadcopter with 8-inch propellers (203.2 mm), the propellers will have a radius of 4
inches (101.6 mm). If the distance between the centers of adjacent motors is 296.87 mm, the
horizontal clearance between the tips of adjacent propellers should be:
Clearance = 296.87 mm − (101.6 mm × 2) = 296.87 mm−203.2 mm = 93.67 mm
This ensures that there is roughly a 94mm gap between the tips of the propellers, preventing
any collisions.
1.4.5 Landing Gear
Landing gears ensure safe take-offs and landings, protecting components, and providing
stability. The landing gear has enough clearance to not disturb the payload dropping mechanism
and is shown in the 2D projection below.
The landing gear perpendicular height is 115.78 mm and the distance between them is 414.01
mm.
Alll 2D plans shown above are attached in the appendix
12
AeroTHON 2024 Team Classified
1.5 Aircraft Performance
1.5.1 Power Required Estimation:
From the datasheet for the motor, we know that the motor can take a maximum of 5s or 21 volts.
However, we will be running it in a 4S configuration as the motor can take that as well.
The maximum power that the motor can take is 960 watts in 4S from the datasheet. However,
from the power estimation, we can see that the actual power required is 90.5 watts at hover and
402.9 watts at maximum desirable thrust to weight ratio.
Maximum Power Required=4*402.9=1611.6 Watts
Hover Power Required = 4*90.5=362 Watts
Energy Required = 362*10/60 =60.33 Watt-Hours
Now if we take the power consumption of other primary components into account, the updated
Energy required is obtained to be 62.4 Watt-Hours as shown in the table below
1.5.2 Power System Battery Selection:
Now, we can see that we need atleast 62.4 Wh just for
the motors and considering all the other accessories and
components, we have decided to go with the Orange
14.8V 6200 mAh 35C 4S LiPo.
Energy = 6200/1000 * 14.8 = 91.76 Wh
Which is more than an estimated 62.4 Wh required.
13
AeroTHON 2024 Team Classified
A PDB is a PCB that has connections to help with
distribution of power from the battery to the motors and
several other components. We have decided to go with
PDB-XT60 with BEC 5v and 12v Power control for our
PDB. This is because of its capability to efficiently
distribute power to the various components we have
decided to go with. Also, since it comes with an XT60
connector, it makes it very easy to use with our battery
and also offers voltage and amperage measurements. It
provides both 12V and 5V which helps to run
components directly. The 5V is mainly required for our
Raspberry Pi 4 which is needed for the video processing.
1.5.3 Endurance Estimation:
The endurance of an aircraft is the time it can fly or the maximum flight time for an aircraft. The
endurance for a quadcopter is dependent on several factors like weight of quadcopter, motor
rating, operating voltage, battery capacity, propeller size and propeller pitch. We have used the
xCopter Calculator from the www.ecalc.ch site. We have the following data for an 8045 Propeller
with our selected motor.
The main data that is of concern here are:
1. Endurance = 10.9 minutes or 0.19 hours
2. Operating voltage = 14.33 V
The range of the quadcopter can be determined as follows:
kv ∗ V ∗ 60 ∗ pitch
Range in miles = ∗ Endurance in hours
12 ∗ 5260
14
AeroTHON 2024 Team Classified
Data:
kv = 1500 kv
Voltage = 14.33 V
Pitch = 4.5”
Endurance = 0.19 hr
1500∗14.33∗60∗4.5
𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 𝑖𝑖𝑖𝑖 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 = 12∗5260
∗ 0.19 = 17.46 miles
1.6 Material Selection
Material selection is important as right materials can enhance a drone's capabilities, making it
lighter, more durable, and more cost-effective. With these considerations in mind, we have
chosen Acrylonitrile Butadiene Styrene (ABS) as material choice for UAV frames and landing
gear due to its unique advantages.
ABS is a versatile thermoplastic polymer known for its excellent balance of strength, impact
resistance, and toughness. With a density of approximately 1.04 g/cm³, ABS is lightweight, which
contributes to the overall efficiency and performance of UAVs. The attached figure shows that
ABS has a good trade off between strength and density. (Refer appendix for chart)
ABS can be easily processed, allowing rapid prototyping and the production of complex
geometries with high precision. It offers a good balance of mechanical properties at a cheaper
price.
1.7 Subsystem selection
1.7.1 Communication system
Transceiver –
The transceiver used of communication between the
drone and the Ground Station is the Fs-iA 10B. This is
directly connected to Pixhawk.
15
AeroTHON 2024 Team Classified
Remote Controller –
The controller will connect to the transceiver aboard
the drone and will be used for manual operation of the
drone and will also be used for activation of
autonomous operation. It will also be used for all
communication with the drone including sending
information to ground station at regular intervals such
as the location coordinates of the drone, status of the
drone, etc.
Flight Controller –
Pixhawk 2.4.8 is the flight controller which will be used for the
drone. It will be connected to the other sensors and
components. It will act as the brain of the drone deciding the
speed of the propeller motors to control the navigation. It will
communicate with the transceiver to execute the commands
received from the Ground Station including movement,
navigation, and other flight controls.
Companion Controller –
The Raspberry Pi-4 8GB microcontroller will
act as a companion controller. The
Raspberry Pi-4 8GB will be directly
connected to the Pixhawk via serial
connection. The main purpose of the
Raspberry is for image processing,
computer vision, object identification and
other visual processes. During autonomous
operation, the Pixhawk will control the drone
based on image processing from the
Raspberry Pi-4 8GB.
16
AeroTHON 2024 Team Classified
1.7.2 Control & Navigation System
ST Micro L3GD20H 16-bit gyroscope –
The gyroscope will be directly connected to the
Flight Controller. It will be used for defining the
orientation of the drone and to measure the angular
velocity of the drone. It’s main use is to help with
maintaining the stability of the drone.
MPU 6000 3-axis accelerometer/gyroscope –
The accelerometer gives us the acceleration experienced by
the drone in the 3 axes (x, y and z). This is used to determine
the velocity of the drone in the 3 axes.
MEAS MS5607 barometer –
Barometer is used to determine the air pressure
experienced by the drone. This air pressure can be used to
determine the altitude of the drone which is used by Flight
Controller for navigation.
Neo 7M GPS Module –
GPS module is used to know the positioning of the
drone along Global latitude and longitude. This is used
by the drone for navigation and for other features such
as Return-to-home, waypoint mapping, etc. This also
contains an inbuilt digital compass.
17
AeroTHON 2024 Team Classified
Idc-681H 25Mw 40Ch Vtx 600Tvl M7 camera –
600Tvl M7 Fpv Camera sends real time video through its
inbuilt transmitter which works in the 5.8 Ghz band. This
camera is mounted on a servo (2D plan of Camera Mount
attached in Appendix) such the camera's facing direction
can be changed on command. While facing down its video
feed will be used for manually coordinating the flight and to
perfectly hover above the payload drop sight.
And it is made to face forward using a servo motor during
obstacles avoidance course and while acting as an FPV
camera. This camera is separately powered by a 1s
160mah LiPo cell.
This Mini AIO FPV Camera is a small, lightweight, and easy-to-use
camera. It features a 5.8G 40CH 25mW video transmitter, an
600TVL CMOS image sensor, and a 2.1mm lens. It is also very
easy to install.
Raspberry Pi cam IMX 519 16MP Camera –
This camera acts as the downwards-facing camera
who’s primary video feed is used for image identification
by the Raspberry Pi for hotspot, object and payload drop
site identifications. It is a high-resolution camera which
comes with a 16-megapixel Mpx sensor. The camera
module is equipped with MIPI CSI interface, which
makes it compatible with microcontrollers such
as Raspberry Pi
18
AeroTHON 2024 Team Classified
1.8 CG and Stability Analysis
1.8.1 C.G. Calculation
Component Count Weight Distance Moment
(g) (mm) (g.mm)
BLDC x4 47.6 0 0
47.6 299.144 14239.2544
47.6 299.144 14239.2544
47.6 420.736 20027.0336
ESC x4 26 101.28 2633.28
26 250.955 6524.83
26 250.955 6524.83
26 327.963 8527.038
Propeller x4 5.2 0 0
5.2 299.144 1555.5488
5.2 299.144 1555.5488
5.2 420.736 2187.8272
GPS module x1 26 77.469 2014.194
Power Distribution Board x1 11 149.022 1639.242
Companion Controller x1 52 180.87 9405.24
Flight Controller x1 40 238.685 9547.4
Camera Forward FPV x1 12 117.164 1405.968
Battery 1S x1 4.65 172.201 800.73465
Camera Down x1 10 285.08 2850.8
Battery 4S x1 533 207.717 110713.161
Servo Motor x3 9 110.561 995.049
9 165.065 1485.585
9 173.655 1562.895
Radio telemetry x1 33 58.31 1924.23
Payload x1 200 212.984 42596.8
Transceiver x1 20 278.07 5561.4
Base Plate x1 99.785 212.984 21252.60844
Top Plate x1 60.133 212.984 12807.36687
Claw x2 13.709 182.686 2504.442374
13.709 243.495 3338.072955
Camera Mount x1 1.451 144.6 209.8146
Arm x4 91.305 73 6665.265
91.305 248.809 22717.50575
91.305 248.809 22717.50575
91.305 346.693 31654.80437
Total 1837.857 394384.5299
394384.5299
𝐶𝐶𝐶𝐶 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 = = 214.5893 𝑚𝑚𝑚𝑚
1837.857
CG is 214.5893mm way from the first motor towards the centre of the drone. Thus the CG is
located (212.984 – 214.589 = 1.60535mm) away from the centre which is within the safety
margins for CG shift from centre of quadcopters.
19
AeroTHON 2024 Team Classified
1.8.2 Stability Analysis
As we can see from the centre of mass in each axes, the drone’s structure is stable with just
13mm deflection in z axis while the other axes are perfectly symmetric.
20
AeroTHON 2024 Team Classified
1.9 Preliminary Computer Aided Design Model
2. COMPUTATIONAL ANALYSIS
FEM for Arm and Drone Frame
A load of 12.14N (max load) is applied on the arms as specified in fig and fixed at the other end
and static structural analysis were carried out and results are compared.
From Total Deformation tests, it is clear that Arm 1 undergoes a larger deformation (max
5.46mm) than Arm 2 (max 1.2mm) for the same load conditions. This is due to the distribution
of the loads due to the Y-shaped configuration present in Arm 2
21
AeroTHON 2024 Team Classified
The Equivalent Von Mises stress tests and Equivalent Strain tests for the straight and Y arm
have been attached in the appendix, which show the minimum and maximum stress and strain
values which prove why the Y arm was a better design.
With the chosen Y arm, the skeleton of the drone is taken and a load of 12.14N(upward) is
applied at the motor hubs and standard earth gravity is applied to take into account the weights.
Since only skeleton is taken in the model the mass of other components is given as uniformly
distributed mass on the top and bottom plates. The deformation and Von Mises stress tests are
as follows. Refer Equivalent strain test results attached in the appendix.
Total Deformation Test
Total deformation represents the frame's displacement under applied loads. The analysis
indicated the quantity and distribution of deformation throughout the frame. We have max and
min deformation of 2.51e-1mm and 0mm. This shows that the frame’s structural integrity is
more than enough to withstand the small deformations.
22
AeroTHON 2024 Team Classified
Equivalent Von Mises Stress Test
The equivalent stress von Mises stress, measures the highest amount of stress that the frame
can withstand. The stress values recorded had maximum and minimum values of 8.202e-1 Pa
and 7.749e-7 Pa. The stress levels are displayed throughout the frame in this analysis with
different concentrations at different points. To guarantee structural safety, it is important to
compare the maximum stress values with the material's allowed limits.
Computational Fluid Dynamics
CFD analysis was done for a single propeller using Ansys Fluent to estimate the thrust produced
and to visualize the flow over the selected propeller.
In the simulation, the propeller is made to rotate at 8915 rpm which is the rpm during hovering
as obtained from ecalc calculations. It rotates in a domain where the flow exits at the top and
bottom and walls at the sides flow so it represents a propeller rotating in an air column. The
propeller blades and hub were assigned no-slip walls, and the surrounding surfaces were set
as walls with appropriate boundary conditions. A transient simulation is conducted with 120
time steps of size 0.5 seconds simulate one minute of flying. The turbulence model used was
realizable k-epsilon with scalable wall functions with turbulence intensity of 5%. Operating
pressure and density were set to mimic Mean sea level conditions. Once the calculation is
done the force report and the contours of flow are obtained from the post processor. As per the
figures below the propeller produces 6.65N of thrust over the calculated time period.
23
AeroTHON 2024 Team Classified
Pressure Contour Velocity Contour
Velocity Streamlines
The force required to get the drone off the ground is 1.905 * 9.81 = 18.68805 N
So each propeller will have to produce at least 4.67 N to hover, and from the CFD analysis, we
can see that we get 6.5 N which is well over the required force.
At a max rpm of 22200, we will get more than the required force to achieve a sufficient T/W
24
AeroTHON 2024 Team Classified
2.1 Optimized Final Design
1) Arm Optimization
The arm designs were changed from a straight arm to a Y arm. The Y arm provided
more structural integrity than the Straight arm as verified using FEA. The Y arm showed
less deformation compared to the straight arm.
The final arm was also made to have a rounded edge and body to ensure its
streamlined.
Straight Arm Y Arm
2) Landing Gear Optimization
The Landing gears were added struts to better distribute the impact force so that it
doesn’t fracture easily
The Final 2D UAV orthographic view is attached to the Appendix
25
AeroTHON 2024 Team Classified
2.2 Recalculation
2.2.1 Detailed Weight Breakdown
Name Component Qty Weight (g) Total Weight (g)
COMPONENTS
ECO II 2807 1500kv BLDC Motor 4 46.9 187.6
ReadytoSky 40A 2-4S ESC 4 26 104
Orange HD 8045 Propeller 4 5.2 20.8
Neo 7M GPS module 1 26 26
PDB-XT60 with BEC 5v Power Distribution 1 11 11
and 12v Board
Raspberry Pi-4 8GB Companion 1 52 52
Controller
Radiolink Pixhawk Flight Controller 1 37.6 37.6
3DR Single TTL MINI Telemetry 1 33 33
Radio Telemetry Transmitter
433MHz 500mW
Fs-ia 10B Transceiver 1 20 20
Idc-681H 25Mw 40Ch Vtx FPV Camera 1 12 12
600Tvl M7
1S LiPo 160mah Battery for FPV 1 4.65 4.65
Cam
Raspberry Pi cam 12MP Down Facing 1 10 10
Camera
LiPo 6200 25C 4S mAh Battery 1 553 553
Screws and Wires Screws and Wires 1 40 40
Brass Heat Inserts Screw attachment 1 30 30
SG90 Servo Motor 3 9 27
Payload Payload 1 200 200
CAD MODELS
Base Plate 1 99.785 99.785
Top Plate 1 60.113 60.113
Rotor Arm 4 91.304 365.216
Payload Dropping Claw 2 13.709 27.418
FPV Camera Mount 1 1.451 1.451
Total 1904.633 g
The weights for the CAD Models have been estimated using the print material in Fusion 360.
Thus, our final estimated weight for the entire drone is 1904.63g.
26
AeroTHON 2024 Team Classified
2.2.2 Aircraft Performance Recalculation
Taking into account the power consumption of all electronic components in the system, the
updated table is as follows which gives us 63.78 Wh to be the required energy
The accessories’ weight and current consumption were added to the www.ecalc.ch site and
calculations were rerun to obtain the following results.
Accessories Weight – 158.6 grams
Accessories Current – 3.445 A
Here we can see that the flight time has changed to 10.7 minutes but we have now also
compensated for the accessories.
27
AeroTHON 2024 Team Classified
Data from the recalculation:
T/W = 2.6
Endurance = 10.7 minutes = 0.17 hours
Hover- Power per Motor = 90.5 W
Total Hover Power required = 362 W
Only the initial assumption to select motor was of a 2:1 T/W, but with our chosen motor and
propeller with all other variables we have a 2.6:1 T/W which is considered stable for delivery
drones.
2.2.3 Final UAV Specifications
S.No. Specification Description
1 FLIGHT MODEL
1.1 Multirotor Type Quadcopter
1.2 Flight Modes Autonomous and Manual
1.3 Payload 200 gm
1.4 Endurance 10.7 Minutes
2 DESIGN
2.1 Physical Dimensions 296.87 X 296.87 X 216 mm
2.2 Weight of UAV 1904.63 g
2.3 Frame Material ABS
3 POWER SUBSYSTEM
3.1 Propulsion Motors EMAX ECO || 2807 1500kv
3.2 Battery 6200 mAh 35C 4s LiPo
3.3 ESC 40A ESCs
4 CONTROL AND NAVIGATION SUBSYSTEM
4.1 Flight Controller Radiolink Pixhawk
4.2 GPS Module NEO 7M
4.3 Communication and Telemetry 2.4 GHz Radio Telemetry
4.4 Payload Dropper Mechanism Servo Motor Actuators SG90
4.5 Companion computer Raspberry Pi 4 8GB
28
AeroTHON 2024 Team Classified
2.2.4 Bill of Materials
Component Name Qty Total Price
(g)
COMPONENTS
BLDC Motor ECO || 2807 1500kv 4 6116
ESC ReadytoSky 40A 2-4S 4 2980
Propeller Orange HD 8045 4 198
GPS module Neo 7M 1 1399
Power Distribution PDB-XT60 with BEC 5v and 12v 1 294
Board
Companion Raspberry Pi-4 8GB 1 7199
Controller
Flight Controller Radiolink Pixhawk 1 14866
Telemetry 3DR Single TTL MINI Radio Telemetry 1 7999
Transmitter 433MHz 500mW
Remote controller Flysky Fi6S and Receiver 1 5295
and Receiver
FPV Camera Idc-681H 25Mw 40Ch Vtx 600Tvl M7 1 1749
Battery for FPV Cam 160 mAh 1S LiPo 1 149
Down Facing IMX 519 16MP Pi Cam 1 3199
Camera
Battery LiPo 6200 mAh 25C 4S 1 6499
Screws and Wires Screws and Wires 1 ~200
Screw attachment Brass Heat Inserts 1 ~100
Servo Motor SG90 3 240
CAD MODELS
Base Plate 1 866
Top Plate 1 455
Rotor Arm 4 4420
Payload Dropping 2 352
Claw
FPV Camera Mount 1 38
Total 64,705
Rupees
The cost for CAD Models have been estimated for each component using an
online 3D printing service from Robu while the other electronic components are
hyperlinked to their respective websites where we plan on purchasing them from.
29
AeroTHON 2024 Team Classified
3. METHODOLOGY FOR AUTONOMOUS OPERATION
Autonomous operation is required for Flight Mission 2 where the UAV is required to survey the
field, detect hotspots and classify and count various objects as well as for Flight Mission 4 where
the UAV must take-off autonomously, survey the field, capture hotspots, find payload delivery
target, deliver the payload and then return to base. MATLAB will be used for Autonomous
operation.
3.1 Autonomous Flight
• Pixhawk flight controller is used as the main controller for the UAV.
• Proper calibration and functioning of Pixhawk is ensured.
• Companion computer Raspberry Pi-4 is connected to the Pixhawk microcontroller via a
serial connection (typically UART or USB).
• Ensure Pixhawk flight controller and the Raspberry Pi-4 are securely mounted on the
frame.
• Required software and necessary items (like operating system such as Pi-OS and other
software such as YOLOv3 orYOLOv4 are installed onto the companion computer.
• Path plan is generated using the geo metric data for mission area, hotspots and object
identifying zones.
• This path plan is integrated using A* path planning algorithm.
• These plans are sent from Raspberry to Pixhawk using MAV Link to start the mission and
the drone takes off and maintains a 15 meter hover altitude.
3.2 Autonomous Object Detection and Counting
• A* path planning algorithm is implemented and will be used to determine the optimal path
to complete mission objective.
• Waypoint is generated based on the path planning algorithm. This is used to map and
cover the entire mission area at a hover altitude of 15 meters.
• After the Take-off sequence, the Pixhawk and Raspberry work in tandem to handle flight
operations and navigation.
• The camera module is used for image capturing which enables YOLOv3 or YOLOv4
which can process and detect the objects at 60 fps by Convolutional Neural Network
(CNN) which can detect large image dataset at high speed with higher accuracy.
• Before going to the CNN, the image pre-processing is done by one of the subsystems
under YOLO called the Backbone sub-system will use CSPDarknet53 or EfficientNet-53
or CSPResNext50 to extract and give the information regarding the frame (such as
Resolution, Time, Location, etc.)
30
AeroTHON 2024 Team Classified
• After the pre-processing is done, the information will be sent to a suitable model to detect
information like the number of objects, colour of the objects and number of similar objects
in the frame.
• After extracting the information from the frame, the frame will be divided into multiple
layers of colour scales. Now each layer is split into number of grids for easy detection.
Number of grids (6x6 or mxn) can be adjusted for our suitable level.
• At the Detection process, each grid is detected by CNN and it will look at the result (from
preprocessing) for that grid like presence of any object in the grid, location of object
present, how much area is occupied by the object, probability of the objects (like tree=0.9,
cat=0.01), etc.
• After the Detection process is done, it will move to post-processing phase which will then
integrate the information from the grids which under went object detection. It will rectangle
in on the frame that the model had detected and we can include some other information
to the rectangle which are object name, color, shape, etc. Also we will include the
timestamp of the frame to our desired location.
• If a hotspot is detected according to the pre-defined criteria, GPS coordinates, time of
identification as well as pictures of the hotspots will be logged into SD card onboard the
UAV.
• Telemetry data will be sent continuously from the drone to the GCS for real-time
monitoring of UAV actions and information.
3.3 Autonomous Payload Dropping
• Payload drop area is detected using YOLOv3 or YOLOv4 and navigation to drop area is
done using path planning (A*).
• Once the payload drop area is detected, UAV will descent to an altitude of 5 meters, and
a releasing mechanism will be activated to drop the payload.
• After the mission is completed, the UAV will rise back to an altitude of 15 meters and will
then go to the finishing point using coordinates and path planning.
• This process can be further fine tuned by tuning the YOLO model for our desired dataset
to increase the accuracy and speed.
31
AeroTHON 2024 Team Classified
4. INNOVATION
- Y Arms
Our drone is employing a Y type split in the arms to distribute the load of stress effectively that
is acting from the landing gear as exhibited in section 2. This further enhances sturdity of the
drone’s structural components.
- Landing gear with Arms
We are also using our arms as landing gears which greatly reduce gross weight of the drone
and give enough ground clearance for the payload dropping mechanism.
The design below shows the idea of having merged the landing gear with the arm, and the
placement enhances stress distribution to the arms.
Drone Arm
Landing Gear
- Payload Dropping Mechanism’s Weightlessness and Simplicity
Our innovative payload dropping mechanism comprises of two L-shaped clamps that are 3D
printed, connected to SG90 servo motors that directly operate the L-shaped clamps to drop the
payload along with minimal place holders which greatly reduce the weight of the payload
dropping mechanism to just 40.48g attached to the base plate.
32
AeroTHON 2024 Team Classified
Our innovative payload dropping mechanism
comprises of two L-shaped clamps that are 3D
printed, connected to SG90 servo motors that directly
operate the L-shaped clamps to drop the payload
along with minimal weight place-holders which
greatly reduce the weight of the payload dropping
mechanism to just 40.48g attached to the base plate.
The 2D plan of L-Clamp is attached in the Appendix.
Base
Servo
Holders
Claws
Payload
4.References
ABS Material Properties
https://plasticranger.com/plastics-used-in-drones/
https://facfox.com/docs/kb/best-materials-for-3d-printed-drone-parts
https://www.grantadesign.com/education/students/charts/
Range Estimation for quadcopters
https://www.airsight.com/learn/airspace-security/drone-fundamentals/drone-capabilities-
endurance-range
Endurance Estimation
https://www.ecalc.ch/xcoptercalc.php
Electronic Components bought from
https://robu.in/
Electronic components CAD
https://grabcad.com/library
33
AeroTHON 2024 Team Classified
Appendix
Rotor Arm
Top Plate
34
AeroTHON 2024 Team Classified
Base Plate
Skeleton Structure
35
AeroTHON 2024 Team Classified
Payload Dropping Servo Claw
FPV Camera Mount
36
AeroTHON 2024 Team Classified
Material Properties Chart
Equivalent Strain Test
37
AeroTHON 2024 Team Classified
Local Pressure Distribution of Propeller
Pressure Distribution about the propeller
Top View of Pressure Distribution
38
AeroTHON 2024 Team Classified
Local Velocity Distribution of propeller
Velocity Distribution about propeller
Top View of Velocity Distribution
39
AeroTHON 2024 Team Classified
Final UAV Orthographic View
Motor Dimensions
40
AeroTHON 2024 Team Classified
