Technical Report

Capstone Design: Maximizing Space in a Cargo Hold

Prepared for Department of Engineering Innovation and Leadership, The University of Texas at
El Paso
by
Christopher Ramirez, Diego Caldera and Diego Martinez
May 06, 2022
 2

Executive Summary
The cargo industry is embracing the expansion of workflow automation to increase the flow of
information from point to point, but lack of existing solutions to evolve into the modern world.
This creates a divide between the digital and physical world which leads to cracks in the system
such as operational inefficiency and space utilization reductions. A vital piece of load handling is
to make sure the correct volume and weight of the cargo are relayed back to the load planners
before it is loaded onto the plane. The data of the volume and weight need to be accurate as it is
critical and a requirement to make sure cargo is arranged correctly on the Unit Load Devices
(ULD) to ensure a safe flight.
 3

Introduction

According to the International Air Transport Association, $6 trillion dollars worth of goods travels by air,
making up approximately 35% of all world trade by value (iata.org). Within the next five years, the air
cargo market is projected to grow more than 5% annually (cargo facts.com). The world and its supply
chains are moving ever faster, but the air cargo industry is not keeping up with demand. (add forecast of
increase in air cargo (e-commerce) demand statistics). Offline and manual cargo-handling processes, as
well as layers of intermediaries, because half-empty cargo holds and losses of millions of dollars. Cargo
handlers and airlines have tried to improve their efficiency but compared to other transportation
industries, air cargo simply is behind.

The Boeing Company is the world's leading manufacturer of commercial jetliners and a service provider
of aftermarket support. The company wants to meet the demands of one of its largest clients by improving
the efficiency and reliability of the cargo-handling process through Digital Models of Cargo (DMC).
DMC is a 3D digital representation of physical cargo via cameras and photogrammetry software. The
solution enables the measuring of volume utilization, monitoring of package integrity, and providing
precise dimensions of cargo that enters an aircraft.

With this groundbreaking integration of DMC into the air cargo industry, stakeholders will be able to
know what the cargo looks like before it arrives at a given destination. The cargo’s placement,
orientation, and how it can be optimized to maximize the space usage in a cargo hold, saving millions of
dollars. Additionally, the tracking of volume utilization per outstation can be calculated using DMC.

For shippers, the integration of Digital Models of Cargo means cheaper, quicker, and fuller cargo flights.
For airlines, it means better volume utilization, less insurance claims, better profitability, and more
control. For Boeing, it means delivering high value worldwide.

This paper reviews the viability of implementing a 3D capturing scanning system within an airplane, the
need for such a device within the air cargo industry, and the processes to validate the concept of the
system.
 5

Project Definition
1.1 Project Aim
The project aims to fulfill the desires of one of Boeing’s largest clients in the air-cargo industry. This
client wishes to improve the gathering of cargo data to better measure volume utilization of individual
Unit Loading Devices (ULDs) that enter an aircraft. This desire comes from observations and
rudimentary studies performed which show that the outstations contracted for cargo handling have not
been maximizing the volume within a cargo hold optimally. Our mission is to solve the need for logistical
data through an innovative solution of capturing 3D digital representations of physical cargo. The
proposed solution is to be incorporated within an airplane.

The main goal of this project is twofold. Firstly, to capture 3D digital representations of physical cargo to
provide an aircraft’s central hub with the configuration of the cargo holds ULDs. Doing so will allow the
central hub to create an optimization plan for identified ULDs. Thus, when the plane is landed, and the
ULD’s unloaded, more boxes can potentially be loaded to these identified ULDs. The result is an
improvement in productivity, better optimization, less resource allocation, and increased utilization of the
aircraft’s capacity.

Secondly, the ability to capture 3D digital representations of cargo and measure the volume utilization of
a ULD will create historical logistical records. These records will serve as a tool to identify which
contracted outstations are loading ULDs to their ideal capacity and which outstations are not. This
performance metric will allow the airline to make informed logistical decisions based on historical data.
To get aircraft to an ideal 80-85% volume utilization.

1.2 Project Objectives

The team must provide a solution that captures data of physical cargo that enters an aircraft cargo-hold.
The data will be used by Boeing’s client to determine unused volume capacity. The team will perform a
proof of concept for this proposed solution. This proof of concept will be aligned with Technology
Readiness Level 3 which is defined as the analytical, experimental, and proof of concept phase. The team
will perform analytical and laboratory studies to physically validate the predictions of separate elements
of the scanning technology as well as its placement. The testing will also include components that are not
yet integrated or representative.
 6

Furthermore, the team will provide supporting information which includes the results of laboratory tests
performed to measure parameters of interest presented by Boeing and its client. Comparisons will be
made to analytical predictions for critical subsystems. Both analytical and laboratory studies will be
required to see if the technological solution is viable and ready to process further through the
development process.

The completed project will help Qatar airways meet its goal of improving volume capacity from 75% to
85% on average per flight route per day. Raising their earnings to an estimated $114,000 USD per flight.
Upon completing the project, the team will also provide Boeing with a detailed report explaining the
viability of incorporating scanning devices within an aircraft.

1.3 Project Requirements

The first requirement for this project is to capture 3D scans of cargo entering an aircraft’s cargo-hold on
individual Unit Loading Devices. The current method of cargo digitization occurs primarily at the main
hub of Boeing’s client. At this site, cargo handlers and load planners are able to optimize cargo capacity
and efficiently assemble ULDs. They do so by using CargoSpectre and SpeedCargo. These systems are
within the airline’s cargo-handling warehouse. At the warehouse, cargo handlers are able to use these
systems to maximize volume capacity to within a few centimeters below the ceiling of the cargo plane.
Qatar airways puts volume utilization high on its priority list because more cargo equals often means
more dollars profited per flight. However, other cargo-handling warehouses do not have these
capabilities. Such warehouses use rudimentary measuring devices in an attempt to maximize volume
capacity. Other warehouses have no standardized method of maximizing volume capacity or do not
prioritize that area at all (Brain Hermsmeyer).

Through several studies, the airline found that certain outstations were more efficient at cargo handling
than others. These studies were done by visually inspecting the cargo once it arrived at the main hub.
Therefore, the airline concluded that if there was a loss of cargo efficiency at the end of the cargo’s
journey - as it arrives at the hub - then there is reason to believe that there is cargo handling inefficiency
throughout the transportation process.

Additionally, we must be able to calculate the volume utilization of each ULD benchmarked against the
maximum space available. Furthermore, our solution must be able to produce viewable outputs of the
 7

aforementioned requirements within 2 hours of a completed scan. These outputs must be exported on STL
files.

As a result, the airline gave the requirement to Boeing to devise a method to capture 3D scans of physical
cargo to better understand the efficiency of cargo handling throughout their clients and contracted
outstations.

1.4 Design Constraints

Every year Boing conducts an annual meeting with its largest clients. During a meeting, Qatar Airways -
who pre-ordered 50 777x freighters - specified that each purchased freighter must have the capability to
capture data of cargo entering the aircraft.

The purpose of the solution being integrated into an aircraft and not fixed at every outstation is simple.
There are outstations that the airline only works with seasonally, once a year, or once every 4 years.
Integrating a solution at every outstation, along with the training required is costly. Furthermore,
investing in an outstation that is only worked with sparsely does not make business sense. However, if the
solution is fixed within an aircraft then wherever an airplane goes, the ability to capture data goes with it.
Regardless of whether the aircraft visits an outstation once a year or once a day.

We found that the best placement for the solution is on the primary cargo door. Due to the placement of
the system on the cargo door, there are various constraints that are given. First is the size. Each scanner
was given an internal constraint by Boeing to be less than 4 x 4 x 4 inches. For weight, each scanner must
be less than 5 pounds in weight. To power the scanning devices, each scanner must be able to be powered
over ethernet and the total drawn energy for the devices must not exceed 50 watts.

1.5 Specifications
This project will explore, prototype, and validate the ability to provide 3D digital representations of
physical cargo via cameras and photogrammetry software. The physical cargo captured will be of the
most utilized ULDs. They typically come in two measurements (96’’ x 126” x 118”) for the PMC ULD’s
and (96” x 96” x 125”) for the P6P ULD’s. Our team will assess the feature, value, and implementation
viability of this capability. We will design a suitable hardware architecture, identify candidate software
solutions, and validate the feasibility of the technology through analysis of desired criteria via a
 8

prototype. The accuracy of the 3D scans of cargo have been given an internal requirement to have an
accuracy of +/- 2 inches.

1.6 Target Values

After analyzing the average volume utilization rate within the air cargo industry, we were able to
determine the percentage that could be optimized per cargo load using our system. Currently, the average
volume utilization rate per cargo load is 60%. Our solution will enable the capturing of dimensions of
each ULD as it enters an aircraft. The receiving outstation will be able to create an optimization plan per
ULD given its current dimensions as well as the maximum dimensions. This will enable cargo-handlers to
possibly place additional cargo on identified ULDs. Our target value of volume optimization is 15% per
cargoload. However, even if we can increase the volume-optimization by 1% there is still large monetary
gain that will be made utilizing our system

1.7 Definition of Key Technical Content

Digitalization is the process of converting information into a digital (i.e. computer-readable) format. The
result is the representation of an object, image, sound, document, or signal (usually an analog signal)
obtained by generating a series of numbers that describe a discrete set of points or samples. The result is
called digital representation or, more specifically, a digital image, for the object, and digital form, for the
signal. In modern practice, digitalized data is in the form of binary numbers, which facilitates processing
by digital computers and other operations, but, digitizing simply means the conversion of analog source
material into a numerical format; the decimal or any other number system can be used instead. [15]

A Unit Load Device (ULD) is either an aircraft pallet and pallet net combination, or an aircraft
container. An airworthy ULD must be structurally capable of restraining the loads and providing
adequate protection to the aircraft systems and structure during flight. ULDs are the only aircraft
parts that leave the control of the airline and return after passing through many unregulated
hands, as most ULD operations are outsourced to ground service providers. Together with the
increasing demands for "shipper-built ULD" from shippers and freight forwarders, it has become
critically challenging for airlines to control and supervise the ULD operations.
 9

Figure 1. Inside air cargo plane

A Unit Load Device (ULD) is used to transport cargo by airplane. This process is also known as
ULD air freight. ULDs can convey any type of cargo, from passenger suitcases and perishable
goods to horses and Formula-1 cars.

Unit Load Devices (ULDs) are the pallets and containers used to safely transport freight in our
aircraft. They allow a large volume of cargo to be bundled into a single unit.

Figure 2. Single Point and Multiple Point Triangulation diagram.

 10

For the purpose of this project, the PMC ULD has been chosen as a basis to run our scans as it is
the most commonly used by the client. [1]

What is lidar?
Lidar (“light detection and ranging”) uses eye-safe laser beams to “see” the world in 3D,
providing machines and computers an accurate representation of the surveyed environment. A
typical lidar sensor emits pulsed light waves into the surrounding environment. These pulses
bounce off surrounding objects and return to the sensor. The sensor uses the time it took for each
pulse to return to the sensor to calculate the distance it traveled. Repeating this process millions
of times per second creates a precise, real-time 3D map of the environment. [2]

Photogrammetry.
Photogrammetry is defined by the American Society for Photogrammetry and Remote Sensing
(ASPRS) as “the art, science, and technology of obtaining reliable information about physical
objects and the environment, through processes of recording, measuring, and interpreting
imagery and digital representations of energy patterns derived from non-contact sensor systems”.
[3][16]

Figure 3. Single Point and Multiple Point Triangulation diagram.

 11

What is Xbox 360 Kinect?

Kinect was launched on 4 November 2010.

Kinect was built to revolutionize the way people play games and how they experience entertainment.
With Kinect, people are able to interact with the games with their bodies in a natural way. The Kinect
sensor lets the computer directly sense the third dimension (depth) of the players and the environment,
making the task much easier.

Kinect’s impact has extended far beyond the gaming industry. With its wide availability and low cost,
many researchers and practitioners in computer science, electronic engineering, and robotics are
leveraging the sensing technology to develop creative new ways to interact with machines and perform
other tasks.

On 1 February 2012, Microsoft released the Kinect Software development kit (SDK) for Windows, which
will undoubtedly amplify the Kinect Effect. The SDK will potentially transform human-computer
interaction in multiple industries—education, healthcare, retail, transportation, and beyond.

The Kinect sensor incorporates several advanced sensing hardware. Most notably, it contains a depth
sensor, a color camera, and a four-microphone array that provide full-body3D motion capture, facial
recognition, and voice recognition capabilities. [4]

STL files.
The STL (STereoLithography) file format is an openly documented format for describing the surface of
an object as a triangular mesh, that is, as a representation of a 3-dimensional surface in triangular facets.
STL is sometimes referred to as "Standard Tessellation Language" or "Standard Triangle Language." The
term used for breaking the geometry of a surface into a series of small triangles, or other polygons, is
"tessellation." The format was developed initially for stereolithography, a form of 3D printing used in the
late 1980s for rapid prototyping. [5]

Cargohold.
 12

Figure 4. Single Point and Multiple Point Triangulation diagram.

Figure 5. Single Point and Multiple Point Triangulation diagram.

The 777F accommodates 27 standard pallets (96 by 125 inches; 2.5 by 3.1 meters) on the main deck. The
industry-standard 10-foot-high (3-meter-high) pallets are accommodated by the large main deck cargo
 13

door. The lower cargo hold has the capacity for 10 pallets, as well as 600 cubic feet (17.0 cubic meters) of
additional bulk cargo. [6][7]

Cargo Handling -
Cargo handling refers to the overall activities in warehouses and logistics centers, such as loading and
unloading cargo, transporting cargo, warehousing, picking, and sorting. The main work in cargo handling
is separated into six tasks: assortment, stacking/taking inventory, transportation, storage (allocation),
sorting, and picking. These tasks greatly affect productivity and quality in logistics. Cargo handling is a
very important part of logistics, and losses in cargo handling directly increase logistics costs. [8]

Figure 6. Single Point and Multiple Point Triangulation diagram.

Cargo Handling
Cargo handling is the segment of the supply chain which processes goods landside in the cargo facility.
From the delivery at the airport of origin until it is ready for loading on the plane, to the unloading at the
destination and handover to the consignee/freight forwarder, many steps are involved with cargo handling
that must be closely followed to ensure shipments are delivered safely and securely.
The process of cargo handling begins with booking and planning shipments, there is much to be done
before the shipment even leaves the facility.
These activities also include the steps handled by freight forwarders as listed below:
 14

 Receive shippers’ requests and check the security status

 Receive shipper freight information
 Plan the routing-direct or consolidation
 Request capacity against forwarder or carrier inventories
 Confirm capacity
 Arrange the pickup of freight
 Picked up from the shipper
The next step in logistics is how air cargo is received and accepted for transport. This is a multi-step
process that includes various stockholders including trucking companies, as well as cargo and ground
handling service providers amongst others. Throughout the shipment process, safety remains a priority for
all parties in the air cargo supply chain. For instance, it’s important to ensure clear and correct labeling
and identification of packages. In addition, the packages themselves must be suited to the content and be
able to protect the goods from any damage. Once the freight forwarding truck arrives at the carrier's
domain and the truck driver has informed the Cargo Handling Agent of their arrival, the carrier domain
can receive the goods. Once the information has been received, and the truck has arrived, an unloading
slot and position will be assigned to the driver. Different slots are given according to needs, such as
dangerous goods, live animals, ULDs, healthcare products, and more.

When accepting shipment as ready for carriage, airport cargo and ground handling personnel must take
several steps to meet the requirements to ship goods by air. First, they must verify if the shipments are
security-cleared. Then they must perform a ready-for carriage check. This entails verifying that all the
information aligns with the actual shipment and ensuring all embargoes and operational restrictions are
applied. Once everything has been checked, the information is validated against the booking and updated.
The primary objective is to ensure the consignments are in compliance with

 Carrier requirements
 Local export rules and regulations
 Rules and regulations of the transit airport(s) and air spaces (if any)
 Import regulations of the destination country

After accepting shipments ready for carriage, airport cargo and ground handling personnel can prepare the
air cargo for flight. The goods in transit must be received and security cleared once again. Goods left on
the aircraft that are in transit are considered transit cargo. Cargo and ground handling services must give
this transit cargo security checks, including x-ray and Explosive Trace Detection (ETD) screening. A
 15

detailed exam of the e-AWB, integrity of the cargo, and piece count is made. Once the pre-plan details are
received from the carrier, a build-up plan must be prepared, which indicates what air cargo is to be built
for flight, and the information is sent to the warehouse.

Figure 7. Single Point and Multiple Point Triangulation diagram.

Now it is time to move the loaded ULDs to a secure flight holding area while being mindful of
all sensitive information such as temperature-controlled and dangerous goods. Ensuring no
flights are delayed, the ULDs can be lined up in order, if it is known, to prepare for ramp
transportation. All ramp safety protocols must be followed. To avoid accidents being mindful of
all ground support equipment during the process of loading and unloading is necessary.
At this time, the control of the air cargo passes from the warehouse operator to the ramp handler.
The transport of goods from cargo terminal to aircraft is a multi-step process best lined out in the
IATA Cargo Handling Manual (ICHM). After following each of those steps explicitly, you
would load the aircraft according to the load plan, making a note of arrangements for special
cargo. Once the aircraft is loaded, any discrepancies must be addressed by updating the
electronic Flight Manifest. It is now that the aircraft can depart, but the cargo loading procedure
is not complete without mailing the flight manifest, loading, and carriage information.
How air cargo is unloaded involves fewer steps for airport cargo and ground handlers than the
shipping and loading process. There are still many rules and regulations cargo and ground
handling personnel must comply with, however. Following ramp safety protocols while
 16

performing tasks in a methodical manner help alleviate accidents and keep aircraft turnaround
times in check.
Once the truck has collected the goods from the carrier’s domain it is transferred to the freight
forwarder hub where it is unloaded and checked. Often before goods are dispatched via a
Forwarder Branch Facility before being handed to the Consignee and then they are finally
delivered to the final customer. [9]

Boeing 777 -

The Boeing 777F is part of the 777 family, in a complete family of jetliners preferred by passengers and
airlines around the world. The market-leading 777 family consists of five-passenger models, with the
ability to fly point-to-point nonstop to bypass crowded and busy hub airports, and a freighter model.
The 777 is available in six models: the 777-200, 777-200ER (Extended Range), a larger 777-300, and two
new longer-range models, the 777-300ER and Boeing 777F.
The market-driven design of the 777 ensures the airplane responds to market needs and customer
preferences. The result is a family of airplanes distinguished by its fuel efficiency, spacious cabin interior,
range capability, commonality, and reliability. The Boeing 777F provides the most payload and range
capability and growth potential in the medium-sized airplane category — all with lower operating costs.
777X has room for 8,131 cubic feet of freight. [10]
What is load planning?
Load planning, also called load optimization, is the process of consolidating multiple shipments and/or
arranging freight to use the capacity of trucks, railcars, pallets, and any other types of containers in the
most efficient way. The main goal is to reduce transportation costs by using fewer vehicles or containers
to move freight. In addition to that, optimizing shipments helps avoid damaging cargo, saves loading and
unloading time, and ultimately, increases customer satisfaction. Moving freight is a very complicated
process with multiple options for interactions between parties. The length of the supply chain can range
from a simple manufacturer-carrier relationship to a multimodal, multi-leg journey, involving a number of
intermediaries (i.e., manufacturer to wholesaler to retailer to freight forwarder) and reloading in several
ports/terminals/distribution centers. Creating an optimal cargo configuration is essential for different
players in the logistics industry — basically, anyone who deals with loading multiple items. [11]

Technology Readiness Level.

 17

Technology Readiness Levels (TRL) are a type of measurement system used to assess the maturity level
of a particular technology. Each technology project is evaluated against the parameters for each
technology level and is then assigned a TRL rating based on the progress of the project. There are nine
technology readiness levels. TRL 1 is the lowest and TRL 9 is the highest.

When technology is at TRL 1, scientific research is beginning and those results are being translated into
future research and development. TRL 2 occurs once the basic principles have been studied and practical
applications can be applied to those initial findings. TRL 2 technology is very speculative, as there is little
to no experimental proof of concept for the technology.

When active research and design begin, a technology is elevated to TRL 3. Generally, both analytical and
laboratory studies are required at this level to see if a technology is viable and ready to proceed further
through the development process. Often during TRL 3, a proof-of-concept model is constructed. Once the
proof-of-concept technology is ready, the technology advances to TRL 4.

Project Context
2.1 General Context
The Boeing Company was started in 1916 and quickly became a world leader in manufacturing and
aircraft design in the 20th century. A major leap forward in aircraft engineering occurred in 1968. Boeing
brought the four-engine 747 to market. The Boeing 747 was the first airplane that was able to transport
full pallets in its cargo hold, which revolutionized the air cargo industry. [12] Today, 90% of the world’s
air freighters are Boeing planes. [13]

Boeing's vision statement is “designed to inspire and focus all employees on a shared future and to
reaffirm that, together, we can meet the challenges that lie ahead.”

Boeing's mission statement is: “To connect, protect, explore, and inspire the world through aerospace
innovation.” Boeing has expanded its client base from national and government contracts to middle
eastern buyers and companies around the world.

Like several other Middle Eastern airlines, Qatar Airways has grown rapidly since its founding in 1993.
The airline's first fleet expansion came with the acquisition of Boeing 747 aircraft. Today, the airline
prides itself on being a world leader in cargo as well as passenger flight services.
 18

In February 2022, the airline agreed to terms with Boeing for the purchase of over 30 777-8 freighters.
The expected value is $20 billion at list prices, making it “the largest freighter commitment in Boeing
history by value”, the airframer says.

Like its 777X passenger aircraft counterpart, the 777-8 Freighter will be the industry's largest twin-engine
cargo jet, according to Boeing. The more fuel-efficient aircraft will have a 118-tonne payload capacity
and a 4,410 nautical-miles range, rivaling its 747-400 predecessor and its Airbus 350 Freighter
competitor. The 777-8 Freighter will have an estimated cargo capacity of 27,056 ft³ / 766.1 m³ [14]. As a
leader in the air cargo industry, Qatar plans to maximize this increase in volume capacity an extra 15%, or
roughly $244 million.

2.2 Literature Review

The Kinect is a structured light (or coded light) system where the depth is measured based on the
triangulation principle. The Kinect consists of three optical sensors: an infrared (IR) camera, an
IR projector, and an RGB camera. The projector emits a pseudo-random pattern of speckles in
the IR light spectrum. This known pattern is imaged by the IR camera and compared to the
reference pattern at known distances stored in its memory.

How Lidar is Used

Cameras produce 2D images of the environment. Lidar “sees” in 3D, a huge advantage when
accuracy and precision is paramount. The laser-based technology produces real-time, high-
resolution 3D maps, or point clouds, of the surroundings, demonstrating a level of distance
accuracy that is unmatched by cameras, even ones with stereo vision. Whereas cameras have to
make assumptions about an object’s distance, lidar produces and provides exact measurements.
For this reason, autonomous or highly automated systems require lidar for safe navigation. The
ability to “see” in 3D cannot be underestimated. Lidar produces millions of data points at nearly
the speed of light. Each point provides a precise measurement of the environment. Compared to
camera systems, lidar’s ability to “see” by way of precise mathematical measurements decreases
the chance of feeding false information from the vision systems to the car’s computer.
Camera performance is also greatly impacted by environmental conditions (e.g., bright
sunlight/glare and darkness) and is, therefore, more susceptible to unpredictable blind spots and
generating false positives or negatives. Whereas cameras are dependent on ambient light
 19

conditions and face challenges with darkness and glare, lidar provides its own light source and
can therefore “see” in all lighting conditions. [2]

PERFORMANCE ANALYSIS OF A LOW-COST TRIANGULATION-BASED 3D

CAMERA: MICROSOFT KINECT SYSTEM
PRELIMINARY TEST RESULTS An out-of-the-box Kinect system was tested for its distance
measurement capabilities. A series of tests were carried out in an indoor environment and the
results are presented below.
The distance measurement quality of 3D range cameras (Chiabrando et al., 2009) and laser
scanners (Glennie & Lichti, 2011) have been shown to be affected by the warm-up time. In a
23.2°C, 883.9 mb, and 36.8% humidity room, a white Spectralon target located approximately
1.1m from the Kinect was observed every 5 minutes over a period of 2 hours. The flat target was
nominally orthogonal to the Kinect and a plane was fitted to the point cloud (Figure 1). During
the first hour of warming up, the estimated normal distance to the best-fit plane changed by 1
cm. Based on Figure 2, it is advisable to warm up the Kinect for at least 60 minutes prior to data
capture. For all the results described, the Kinect was turned on at least 90 minutes prior to data
capture.

The Kinect system has demonstrated the potential to be used for engineering applications. At a
long-range, it may not be as accurate as a terrestrial laser scanner, photogrammetric system, or
structured light system, but for the cost and portability, it is delivering fairly high geometric
accuracy at close-range. The Kinect was used to capture point clouds in a series of tests. Since
this is a triangulation-based system, it does not exhibit data distortions that would otherwise be
identifiable in TOF systems, of which the most pronounced range error is the internal scattering
experienced by most TOF 3D range cameras.

Calibration of Multiple Kinect Depth Sensors for Full Surface Model

Reconstruction
The main function of the scanning system is for the reconstruction of the full surface model of an object.
In this work, we build a four-Kinect system that the Kinect range sensors are positioned around the target
object. Each Kinect is responsible for capturing a small local model, and the local models found will be
 20

combined to become the full model. To build such a system, calibration of the poses among the Kinects is
essential.
3-D scanning is important in many scientific and industrial applications. Nowadays, 3-D scanners are
consumer-grade products which are available at low cost. In general, if we want to capture the full surface
model of an object, you may use a single Kinect to move around the object to obtain the result.1
However, this task is relatively difficult to be handled and carried out perfectly. The method here is a
multiple Kinect approach in that each sensor is set to obtain one face of the whole model and these small
models are combined at the end. The drawback of using multiple Kinects is that (i) it is a relatively high
cost, and (ii) the system requires calibration beforehand. Carrying out calibration is necessary especially if
we need to transport the system from place to place.

Nowadays, there are many range sensors available in the market, which are convenient tools for
performing conventional computer vision tasks such as 3-D structure reconstruction.2–4 The Microsoft
Kinect,5 which is a popular and consumer-grade range sensor, is first designed for games and
entertainment purposes. Since it is inexpensive and accurate enough, it has gradually become a popular
tool for computer vision researchers.

The Kinect sensor together with some open source libraries, e.g. Kinect Fusion9 are low-cost solutions for
obtaining the 3D model of an object.
In order to scan a 3-D object for its full 360◦ views, it requires at least three Kinect sensors aiming at the
object from different angles. To achieve that, we also need to calculate the relative poses among these
sensors. Therefore the model segments from each Kinect can be integrated into a complete 3-D model.
The computation of the pose of a pair of Kinect sensors is performed first. After obtaining the poses of all
pairs of neighboring Kinects, we will have enough information to stitch the 3-D surfaces to become the
full model. The calculation of the extrinsic parameters between a pair of Kinect sensors is the most
critical step in the whole system. Its accuracy determines the quality of the combined model. We used 4
Kinect (version1) sensors and an A3-sized 9x6 checkerboard pattern in the experiment. We put the
checkerboard in front of the sensors and then capture the depth snapshot by each pair of Kinects one after
the other. The 3-D data captured, together with the pattern on the board, were used in the normal equation
to find the result. [17]

Value of Air Cargo

 21

Air cargo is a trade facilitator that contributes to global economic development and creates
millions of jobs. The global economy depends on the ability to deliver high-quality products at
competitive prices to consumers worldwide. Air cargo transports over $6 trillion worth of goods,
accounting for approximately 35% of world trade by value. [19]
Within a 24-hour period over 80 thousand flowers, 20 million parcels, and 1.1 million
smartphones are shipped totaling over 18.6 million dollars.

2.3 Review of Earlier Projects

Airlines have a difficult time keeping track of their ULDs, containers, and pallets throughout the cargo
handling process. This is due to the speed of the process and the continuous loading and unloading of
cargo. Typically, a cargo's location is determined by a freight status update message which is received
when cargo reaches a critical stage along the shipping trip, such as customs clearance. These
communications are frequently sent manually by operators(managers). However, they are often not sent at
all. As a result, there is an estimated 5-10% ULD missing rate at any given time said a UPS representative
during our visit.

During our visit to the Boeing facility in Everett, we found that the Connected Digital Systems division
has been working to improve this problem within the air cargo industry. Their team has considered the
employment of RFID (radio frequency identification) tags, cellular, GPS (global positioning system), and
Bluetooth Low Energy(BLE) devices to offer real-time tracking of ULDs. This will enable Qatar Airways
to see online where any ULD is located. When ULDs are assembled, for example, overhead smart
antennae identify the contents automatically, and when the ULDs move about the warehouse, readers at
crucial points update the ULD's location. This same concept will be applied when ULDs enter the cargo
hold of an aircraft. This is key as certain ULDs will stay on the plane and others rushed off depending on
their final destination. If a ULD goes missing, it can be tracked to a precise location.

The implementation of this technology improves the digitization and automation of the cargo-handling
process. The incorporation of our solution along with the tracking system mentioned, will not only tell
you where a ULD is located it also tells you what the assembly of the ULD looks like. The technology
can be incorporated throughout the cargo handling process and will be connected to the company's
Internet of Things.

Lidar Used
 22

Lidar sensors are a key component in autonomous vehicles, providing a high-resolution 3D view of their
surroundings. Lidar enables autonomous vehicles to “see” by generating and measuring millions of data
points in real-time, creating a precise map of its ever-changing surroundings for safe navigation. Lidar’s
distance accuracy allows the vehicle’s system to identify and avoid objects at up to 300 meters in a wide
variety of weather and lighting conditions. [2]

Kinects Used in the Cargo Handling Industry

A small local company is tackling one of the most vexing problems for the warehousing and
transportation industries: measuring the dimensions and weight of a piece of freight, quickly and at a low
cost. It says it has found the solution — in video games. Cargo Spectre, founded about five years ago, has
adapted video game technology for a system of scales, cameras, and 3D scanners to measure and record
within seconds the weight and dimensions — height, width, depth — of freight, including pallets loaded
with goods and heavy machinery. This product, known as a dimensioner system, can make and record
measurements in a fraction of the time it would take to take such measurements by hand, at one-tenth of
the cost of similar devices marketed by competitors, according to the company co-founder and CEO Jason
Joachim. The Cargo Spectre system takes about three seconds to complete measurements and record the
data, which is uploaded to the cloud and instantly available to employees and clients. “This system has
worked well for larger customers that have facilities all over the world,” Joachim said. Cargo Spectre’s
beginnings reach back to 2015, when Joachim and his brother, Jeremy, now Cargo Spectre’s chief
technology officer, were working for their family’s business, World Trade Distribution, which offers
freight, cargo, and warehouse services at its container freight station in Houston. The company’s freight
processing system desperately needed updating, Jason Joachim said, and they set out to find an affordable
solution for weighing and measuring cargo electronically. At that time, he said, electronic dimensioning
systems cost upwards of $130,000, a hefty price tag for a medium-sized company. With his brother, a
software designer, and a few video game enthusiast friends, Jason Joachim sought to develop a low-cost
solution to the dimensioning challenge. Rather than developing the technology from scratch — a very
expensive proposition — Joachim and his team began searching for off-the-shelf equipment they could
adapt. They found it in Xbox Kinect sensors, an add-on for the Xbox 360 video gaming console produced
by Microsoft. The sensors contained a 3D camera, which could be adapted for scanning irregularly
shaped pieces of freight and computing its precise dimension. Selling for about $300 when new, Kinect
provided an inexpensive platform on which to build a new dimensioning system. The team then wrote the
software that records, calculates, and transmits the data in a matter of seconds. After putting the system
into operation at World Trade Distribution, the development team formed Cargo Spectre in 2016. The
 23

name grew out of the early efforts to develop the technology employing the Kinect. “The first time we
saw ourselves underneath the sensors in 3D space we looked like ghosts or specters,” Joachim said.
Cargo Spectre operates on a hybrid hardware/software business model. The company sells the equipment,
which costs about $5,000 installed. This price compares with about $50,000 for a dimensioner system
offered by its closest competitor, Joachim said.

Customers also pay a $600-per-month fee for the use of Cargo Spectre’s proprietary software, which can
be integrated into customers’ software, allowing the data to flow seamlessly into record-keeping and
management systems.

Precision equals money Precisely measuring the dimensions of an object being shipped is important to
determine the price paid for shipping. Cargo Spectre’s customers comprise freight forwarders, which
transport goods between two destinations; common carriers, which transport goods on regular routes at
set rates; and freight brokers, which are contracted by shippers to be liaisons between the shipper and a
motor carrier to facilitate the movement of cargo. All these types of businesses are affected by cargo
rates. Cargo transport rates are paid based on the dim weight of a piece of cargo, that is the amount of
space a package occupies in relation to its actual weight. Joachim said this is where precise dimensioning
of pieces of cargo becomes important, especially when the product to be shipped is irregularly shaped.
For example, a freight forwarder measuring such a piece of cargo by hand could underestimate that
piece’s dim weight, causing the freight forwarder to receive less money than it should. While the
inexpensive Xbox Kinect helped Cargo Spectre keep its hardware prices low in the company’s early days,
that arrangement came to a halt when Microsoft canceled the production of Xbox Kinect in 2019. But
what looked like a setback turned into another opportunity for the company. Microsoft employed the
same scanner technology for gaming to create the Azure Kinect, which uses artificial intelligence sensors
that have broad commercial applications far beyond video games. As a result, the Azure Kinect proved an
advance for the Cargo Spectre system Joachim said. “We have seen tremendous gains in accuracy,
stability, speed, and AI capability,” Joachim said. “Microsoft has been very responsive to our needs with
this new sensor versus one built for gaming.”

Cost v. precision Cargo Spectre’s low-cost approach is not shared by some of its larger, more established
rivals. For example, Cubiscan, of Farmington, Utah, offers a wide variety of large-scale dimensioning
equipment, most of which are sold for many times what Cargo Spectre’s dimensioners cost.

-Speed Cargo
 24

CARGO EYE, developed with Microsoft's cutting-edge Time-of-Flight sensing technology, captures
imagery of incoming cargo and calculates accurate dimensions based on cargo type, size, and packaging
in real-time, while seamlessly integrating the data into existing warehouse management systems. The
solution also feeds data into CARGO MIND to accurately build load plans, monitor, and promote data
compliance, and identify damaged cargo to reduce potential claims.

“The technology has the added advantage of significant cost control through manpower savings which
could reach 3,720 hours a month, and in minimizing leakage and optimizing offload recovery by up to a
third, which is a result of improving customer satisfaction by automating accurate data receipts. Clients
will also benefit from highly transparent and accountable billing as well as much faster response times for
freighter charter quotations.”

“Industry estimates suggest world air cargo traffic will more than double over the next 20 years,
expanding to $578 billion revenue tonne-kilometer (RTK) by 2039; the industry has to be digitally ready
to take advantage and secure increased market share,” said Drew. The technology is adopted in other
forms of cargo/handling and transportation. [20]

2.4 Stakeholder Definition

The stakeholders for our project consist of several key players within the cargo handling process as well
as an airline and design team. A key stakeholder that will aid the integration of our solution is the
Connected Digital Systems division within Boeing. Currently, the team has worked on other aspects of
Digital Models of Cargo such as a tracking system for ULDs, improvements to the plane’s external wi-fi,
and integration of various devices to the company's Internet of Things (IoT).

The next key stakeholder is Boeing’s client Qatar Airways. This airline has established itself as one of the
world’s leading companies in air cargo services and customer peace of mind. The airline is the driving
force behind the desire for this solution. They are also the end-users of the system. The full development
of the proposed solution will enable them to make better operational and logistical decisions in the air
cargo section of their company.

Other stakeholders for our project include the outstations that are contracted by Qatar Airways. These
outstations are responsible for a significant portion of the cargo-handling process. Although Qatar
Airways itself performs cargo handling within their hubs efficiently, they have minimal monitoring of the
 25

performance of its outstations. With the integration of our system, these outstations will now be measured
against growing logistics records to measure efficiency at volume utilization. Additional stakeholders
include shippers, consignees, and freight forwarders who will be affected by the implementation of DMC.

2.5 Definition of Stakeholder Needs

Our team was able to speak with representatives from each of the stakeholders and learn about their needs
throughout the project. The Connected Digital Systems division within Boeing wants us to explore,
prototype, and validate the ability to provide 3D digital representations of physical cargo. In addition,
they want us to evaluate the capability's feature value and implementation viability. Finally, they are
counting on us to create a suitable hardware architecture, identify potential software solutions, and test the
technology's practicality using a prototype to analyze desired criteria. Their team will take the findings
from our proof of concept phase and continue the system’s development into an aircraft. This division is
relying on our results to help finalize a fully developed solution ready for launch by 2027.
Qatar Airways has the end goal of receiving a final product from Boeing that can capture data that will
enable the measuring of volume utilization as well as the capturing of dimensions per ULD.

Methodology
3.1 Definition of Tasks
Throughout the year, many tasks were completed in order to meet the project's objectives. The project
was divided into two main stages. The initial step was to define the problem and come up with the best
solution. The second was to demonstrate the concept's viability in light of the project's scope.

Figure 8. Spring sprint plan.

 26

The first task that was completed was to create a sprint plan for the first major process. This sprint plan
was centered on the project statement, sprint backlog, and the major objectives within the project. The
project statement includes a summary of the project, the need for a solution, and the impact of the
implementation of the solution. The sprint backlog which is divided by weeks includes a collection of
work items that our team finished throughout the project. The major objectives within the sprint plan
include categories for the various components necessary to complete the project goals.

The next task was to create a SWOT analysis. The SWOT analysis was created to identify the strengths,
weaknesses, opportunities, and threats (SWOT) between volume-optimization methods used within the
air cargo industry and other transportation industries. This was accomplished by attending site visits to
cargo-handling facilities as well as conducting interviews with various experts. This gave us a thorough
understanding of the variables and techniques that must be considered when volume optimizing. From our
SWOT analysis, we discovered that there are three categories that are used in volume optimization. The
first segment we named Standard Operating Procedures - any solution that requires step-by-step
instruction on how to perform a technical, repetitive, process such as the assembly of a ULD. The second,
computational systems - a standalone device that can obtain and process dimensions mathematically. The
third, manual data - a user who would have direct interaction with a device such as a handheld 3D scanner
or a measuring device.

With a holistic understanding of the nuances that go into volume optimization, the next step was ideation.
After narrowing down our ideas into three segments we created a mind map. A mind map is a graphical
representation tool that can be used to assist in structuring information, help analyze data better, and help
generate ideas. Our mind map allowed us to organize our thoughts and make connections between
possible solutions.
 27

Figure 9. Ideation Board of 58 possible solutions.

The aim during ideation was to devise a multitude of divergent ideas based on the analysis that we
performed. First, we created an ideation board. The primary purpose of an ideation board is to brainstorm,
highlight as well as collect ideas, and create designs. During this task, each team member was assigned a
certain category to brainstorm ideas on. Each member then added various ideas related to improving
volume optimization within a 15-minute period. We then switched categories for another 15 minutes until
each team member ended up with the category he started with. This method was conducted various times
throughout the ideation phase.

The next task was for us to create a Pugh chart. A Pugh chart is a tool used to compare multiple solutions
within a category against each other. In contrast to discussing the pros and cons of each solution, the Pugh
chart gave a comprehensive overview of the client's needs, the intended goals for our solution, and the
best solution as well as the category to pursue. The next task after selecting a solution was to create a
SWOT analysis to choose the placement of the system within an aircraft. Our next task was to develop a
design proposal based on the best-rated solution from our Pugh chart and SWOT analysis. We then
presented our proposal to the Connected Digital Systems division within Boeing.
 28

Figure 10. CAD Model of our system showing the ULD, cargo door, and field of view of cameras.

After our design proposal was accepted, we then began to validate our solution by first making an
AutoCAD model. This model was made using the dimensions of the cargo door which is the placement
point of our system within the aircraft. The angle of the cargo door opening was determined by
conservatively analyzing the opening of a door during the loading/unloading phase. We also added a
PMC ULD to the model to represent cargo scanning at the capture point of our system. Then we included
our 3D scanning devices (Microsoft 360 Kinects) with each projecting its most effective field of view
(FOV) ( please view Figure 11.) provided by the specifications. In the AutoCAD model, we placed each
scanning device in a position to maximize its field of view while capturing complete coverage of the top
view. The purpose of capturing the top view (Please view Figure 12.) is because the other views can be
calculated mathematically. For example, if we take a preliminary scan from the scanning devices to the
floor and then subtract that distance from the height of the cargo, we receive the depth dimension.
Additionally, since the cargo that we are focusing on is prismatic, we can conclude the dimensions for the
side views by capturing the top view and the depth of the cargo.
 29

Figure 11. CAD Model of our system showing the field of view of cameras.

Figure 12. CAD Model Top View of our system showing the field of view of cameras.
 30

Figure 13. Model of our system showing the ULD (bottom), camera rigs (top), and 360 Kinects (top in
black) within the mock-up space.
 31

The next task was for our team to find software suitable for our hardware solution. We completed this
task by researching programs, reviewing their capabilities, downloading them, and choosing the program
that enabled the best functionality of the hardware to meet the project requirements. Next, we began
testing the compatibility and capabilities of our hardware and software solution to provide desired results.
Our next task was to create a mock-up space (Please view Figure 11.). We based our mock-up space on
our AutoCAD model. To accomplish this task, we assembled large boxes into the shape of a PMC ULD.
We then built two camera rigs to mimic the scanning device's placement since the devices are positioned
on the Y-axis. Although it appears that three rigs would be necessary, we used two: one for the furthest
edge and one for the middle. This is due to the furthest edges of the scanners mirroring each other on the
X-axis. After completing a scan using one rig, we flipped its placement to reflect the edge on the opposite
side of the cargo door. The positions of the individual scanners were determined using the dimensions
from our AutoCAD model. The angle of each device was calibrated using a digital leveler to ensure that
the field of view on the computer model matched our mock-up.

Our next task was to create a mock-up space (Please view Figure 13.). We based our mock-up space on
our AutoCAD model. To accomplish this task, we assembled large boxes into the shape of a PMC ULD.
We then built two camera rigs to mimic the scanning device's placement since the devices are positioned
on the Y-axis. Although it appears that three rigs would be necessary, we used two: one for the furthest
edge and one for the middle. This is due to the furthest edges of the scanners mirroring each other on the
X-axis. After completing a scan using one rig, we flipped its placement to reflect the edge on the opposite
side of the cargo door. The positions of the individual scanners were determined using the dimensions
from our AutoCAD model. The angle of each device was calibrated using a digital leveler to ensure that
the field of view on the computer model matched our mock-up.

The next task was to calibrate the Kinects. We completed this task by first measuring the distance
between each scanning device (Kinect) using a measuring tape. Second, we inputted the distances into the
software code script within our computer program (Kscan). Third, we numbered each device based on its
position and view. Fourth, we performed a preliminary scan of each Kinect’s view. Fifth, by knowing the
view of each numbered Kinect, we meshed the scans in the software manually to replicate the physical
object. Sixth, we took the dimensions of the scanned object using the software. We then adjusted the
meshed scan’s digital dimensions until they matched the object’s physical dimensions. Finally, once the
 32

digital measurements of the object matched the physical dimensions, our calibration was complete (Please
view Figure 14).

Figure 14. CAD Model of our system showing the meshed model.

The next task was for us to begin testing the 360 Kinects in the area of distance capabilities. We
completed this task by determining control as well as variable conditions. We increased the variables
based on the requirements given to us by Boeing. The control values were determined using information
that we received in the cargo-handling processes along with aircraft operating conditions. Additionally,
we compared our findings with the manufacturer specifications for the 360 Kinect.

In calculating the volume utilization percentage, we first used the dimensions of the fully optimized PMC
volume. The dimensions of the fully optimized PMC are as follows. Having a length of 125 inches, a
width of 96 inches, height of 118 inches. The PMC is fully optimized by mimicking the curvature of the
plane. For one side of the PMC, it will be a height of 96 inches from with a width of 48 inches. With
these dimensions, we get a base volume of 1,284,000. With this controlled number which is the 100%
volume utilization, we can use it to compare to the rest of the ULD’s. We created a coding as a function
of calculating the volume utilization in comparison to the PMC.

The next task was for us to perform a cost-benefit analysis. We completed this task by researching the
amount of payload a Boeing 777 can carry. Next, we communicated with our mentor to receive the
 33

average cost their airline client charges per kilogram. Upon gaining these two pieces of information, we
created a table for the value of a payload at 1%,10%,25%,75%, and 100% utilization to determine the
monetary value for the airline at each rate. After determining the potential losses and gains for the airline,
we began determining the cost of our system. The costs included hardware, software, installation, and
maintenance. Upon obtaining this information, we calculated both return on investment and the payback
period. The next task was for us to revise the results and refine our design. Finally, we communicated the
results back to Boeing.

3.2 Relevance of Tasks to Desired Outcomes

Our first project goal was to define our problem in a clear and concise manner. The second goal was to
create an implementation plan in order to place our system within a Boeing 777 aircraft. The third project
goal was to propose a solution that is able to capture 3D representations of cargo and meet the specified
requirements as well as constraints. Our fourth project goal was to create a digital model of the system
and a prototype to test that model. Our final goal was to refine and validate our model to generate the
desired results. We completed tasks that allowed us to gain a better grasp of the project in order to
accomplish the first project goal. These tasks included creating a sprint plan, conducting interviews with
industry representatives, attending site visits, and analyzing the cargo finances of Boeing’s client. To
meet the second goal, we performed tasks that allowed us to understand how volume optimization has
been implemented in other industries as well as analyses to determine the best placement point of our
system. These tasks included using strategic planning and strategic management techniques,
brainstorming through mind maps, conducting interviews with Boeing engineers in the Connected Digital
Systems division, and synthesizing ideas through a Pugh Chart. To meet our third goal we combined the
findings from our analyses and proposed a feasible solution for Boeing. Our fourth goal was met by
performing tasks that included utilizing computer-aided design software, calibrating scanning devices,
creating scripts of code, and constructing a mock-up space to validate our graphical model. The final goal
of our project was met by performing tasks to revise and improve the variables within our system until the
desired results were produced. These tasks also included re-calibrating, adjusting our camera rigs, as well
as replacing malfunctioning equipment.

3.3 Progression of Work

The first objective that we focused on was defining Boeing’s client’s problem. To do this we
created a sprint plan. The first item that we had to dissect was the project statement. Through the project
statement, we defined that the target area is cargo passing through the outstations of various
 34

intermediaries. We interviewed cargo handling crew members during a tour of United Parcel Service
(UPS) Air Cargo El Paso. At the facility, we have explained the Standard Operating Procedures (SOPs)
for aircraft cargo loading and viewed their execution. Michael Shillito, UPS operations supervisor, was
flown in from Louisville, Kentucky to assist in our tour. Michael has over 15 years of experience in the
industry and is responsible for the cargo handling of over 120 flights of cargo per night at the UPS station
in Kentucky. During the tour, we recorded the live loading of two aircraft; a Boeing 747 and an Airbus
A350. At the UPS site, we interviewed two Aircraft Load Planners (ALPs) who explained the issues that
they have witnessed throughout their experience. Furthermore, the Aircraft Load Planners explained to
what degree weight affects loading an airplane. During the visit, our team was able to learn the methods
of loading a plane, the software necessary to position cargo in certain areas, the importance of cargo
weight in certain areas of the plane, the issues that arise during loading operations, what works well, what
needs improvement, AMJ, Load Plans, how cargo is loaded and locked into position. It was important to
visit the air cargo facility because we could not see the intricacies that go into air freight from mere
research alone. Additionally, many of the details of air freight are not shared publicly and require security
clearance due to the events that occurred during 9-11.

After having a complete understanding of the cargo-handling process and the roles of intermediaries as
well as outstations, we moved on to the ideation phase. This included our team using an ideation board
and mind map\to display our ideas. As we learned more about the variables that cause the problem, we
continuously added new systems as well as methods to solve it. From the ideation process, we discovered
that a computational system is the best method of solving the problem through our Pugh chart. We found
that the best solution would be integrating Speed Cargo and its components into an aircraft.
 35

Figure 15. Visual representations of SpeedCargo’s system.

This system uses photogrammetric and lidar devices, the Microsoft Azure Kinect, to capture the scans of
objects. Additionally, their system can calculate volume utilization. Then we considered the technical
specifications of the SpeedCargo system and methods for its proof of concept in a cargo hold. After that,
we conducted a series of analyses as well as interviews and found that the best place to implement the
system would be on the cargo door of the 777.

3.4 Considered Alternative Solutions

Figure 16. Pugh chart to determine the best possible solution.

We used our pugh chart to determine two solutions that best achieve the project goals. The two solutions
are computational systems that can be implemented into a plane. We explored alternative solutions that
could help us achieve our project objectives if the initial solutions are not sufficient. The first alternative
solution considered was a change to the SOPs. Standard Operating Procedures were considered because
we found them to be the easiest to implement. We found that Boeing’s client is exceptional at maximizing
cargo space on an airplane. The solution of training personnel at the outstations contracted to Boeing’s
client is feasible. Training will ensure that cargo matches manifest descriptions, is handled correctly and
does not encounter conditions that might lead to damage. Additionally, the training will instruct ground
handlers on how to prepare ULDs to fill the cargo hold within centimeters of volume capacity. However,
this process is “labor-intensive and requires a long paper trail that must be precise all while trying to turn
cargo around as rapidly as they can.” We also considered several handheld measuring devices. Each
device requires the training of personnel to capture the data necessary to complete the desired objectives.
 36

The EinScan, Artec3D, Revopoint, Clicklog, and MagicPlan were considered. The devices use laser
technology, lidar, and photogrammetry. Handheld technological devices were considered because other
rudimentary handheld tools are used to measure and capture data currently

Validation
4.1 Plan for Evaluation of Outcomes
Four items will be evaluated: the accuracy of the dimensions that our system produces, the accuracy of
our mock-up space to match the assumptions of our graphical model, the ability for our system to capture
a 100% scan of the top view of a ULD, and the results from our cost benefit analysis. The accuracy of our
system will first be evaluated by measuring the dimensions of an object using a tape measure and then
scanning that object using the Kinects. After the scan has been completed, and calibration has been
finalized, we will compare the dimensions obtained from our system with those using a tape measure. The
accuracy must be +/- 2 inches to meet the requirements from Boeing. If any dimension received using our
system, along the X, Y, or Z-axis is off by more than +/- 2 inches, we will recalibrate the devices,
document the changed variables, and test again until we meet the requirement.

The graphical model that we made using the specifications of the scanning devices, cargo-door, and
ULDs will be used to create our mock-up space. We will do so by constructing camera rigs to simulate
the Y-axis that the cameras are on. Then we will construct a ULD to be scanned. If we are unable to
replicate the positions of the scanning devices or the dimensions of the ULD and cargo door then we will
have to scale the mock-up space.

The ability for our system to capture a 100% scan of the top view of a ULD will also be evaluated. We
will do so by comparing the scanned image to the physical dimensions of the ULD. If there are gaps in
the scanned image, we will determine that the system has failed to capture a 100% scan of the top view of
the ULD. We will then recalibrate the system by positioning the scanning devices until they capture the
entire top view.

The results from our cost benefit analysis will be evaluated based on our ability to provide a profitable
solution with a positive return on investment. If our solution’s implementation does not provide a positive
ROI, we will reevaluate our analysis and provide alternative economic solutions.
 37

4.2 Execution of Plan

The first objective that we focused on was creating a digital model of the system and a prototype to test
that model. To do this we first had to discover the smallest ULDs, the most common ULDs, as well as the
configuration of the most volume optimized ULDs that Boeing’s client uses. We did this so that way we
can see if our proposed solution can produce outputs given these three categories. During this discovery
phase, we discovered that most of the ULDs that the client uses are prismatic. Thus, we learned that we
can scan one side of the ULD and mathematically calculate the other sides. For the purpose of our testing,
we chose the top view. Then we obtained the dimensions of the cargo door of a Boeing 777. Finally, we
gathered the technical specifications of the 360 Kinect. (Note: we used an Xbox 360 Kinect for our
testing, not an Azure Kinect due to its unavailability during the coronavirus pandemic.)

After gathering the dimensions and technical specifications of the ULDs, Kinects, and cargo door, we
created digital models of them on AutoCAD. The purpose of this graphical representation was for us to
see how many Kinects would be required to take a scan of the ULD. We tested different placement points
as well as different amounts of Kinects until we received complete coverage of the top view of the ULD.
Through our graphical model, we found that our system would require (9) 360 Kinects. (3) Kinects were
grouped in the center of the cargo door. Then we placed (3) Kinects close to the furthest edge of the door
and mirrored the final (3) on the opposite side. This was due to the greatest field of view for the Kinects
to capture the ULDs. This position also considers the possibility of the client loading ULDs in different
rotations. We performed the same method using the technical specifications of a Microsoft Azure Kinect
and found that our system would only require 6 Azure Kinects (Please view Figure 17). This is due to the
expanded field of view for the Azure Kinect (Please view Figure 18).
 38

Figure 17. CAD Model of our system showing top view for Microsoft Azure.

Figure 18. CAD Model Showing field of view comparison between Microsoft Azure/360 Kinect.

Our next objective was to validate the graphical model. To do this we created two camera rigs. The
camera rigs represent the Y-axis planes that the cameras are positioned on based on the top view of the
cargo door. Although there are three planes, two of the planes mirror each other. Therefore, we created
the middle plane and the furthest right plane using the distance dimensions from the graphical model.
Once the planes we created, we mounted the Kinects onto the camera rigs to replicate their placement
 39

points. Then we used a digital leveler to ensure that each Kinect was at the exact degree shown in our
graphical model. After that, we connected the devices to our computer and ran them on Kscan, a 3D
scanning software. The software requires proper calibration of each Kinect to produce an accurate digital
model of the scanned object. Additionally, we created a ULD to scan and test our system. After we
performed a scan of the top of the ULD we compared the software’s dimensions to the dimensions we
retrieved using a tape measure. We recalibrated the devices until the dimensions matched and our
graphical model was validated.

The final objective of our project was to create a cost-benefit analysis. Upon doing so we gathered
information from our mentor and through research, the team learned that Qatar Airways is losing roughly
$600 million annually. We also learned that with our system, Qatar Airways could meet their 10%
payload increase and increase their annual revenue by approximately $244 million. Upon learning the
benefit that our system could benefit the airline by this large margin, we then decided to create a
breakdown of the system's cost. We calculated $42,300 for hardware, $3,154 for back-up hardware,
$3,000 for software, $25 million for research & development, $4,897,500 for installation, and $364,000
for maintenance. All these individual parts totaled up to roughly $37 million in cost. We were then able to
calculate an ROI of 662% with a payback period of 2 months and completed our cost-benefit analysis.

4.3 Met Objectives

The first objective of creating a graphical model of the system and a prototype to test that model was met.
We completed this objective by first gathering the dimensions as well as technical specifications of a 777
cargo-door, a PMC ULD, and 360 Kinect. We then translated this data to AutoCAD and assembled them
on the software to accurately reflect a ULD being scanned on the top view by our implemented system.
Next, we created a physical mock-up space using the graphical placement positions of the cargo-door,
ULD, and 360 Kinects. Finally, we calibrated the Kinects and compared the dimensions from our scan to
the physical dimensions received on a tape measure. Once the dimensions matched, our graphical model
was tested and validated and this objective was met.

The cost-benefit analysis was a systematic method of quantifying and then comparing the total costs of
the implementation of our system to the total expected benefit. We completed this objective by gathering
financial data from the cargo industry through interviews with the Connected Digital Systems division
within Boeing and conducting research. We then applied this data to several formulas to calculate the
costs, benefits, as well as return on investments. The results were positive returns on investments as well
 40

as a profitable solution. Once all calculations were performed and documented, our cost-benefit analysis
objective was met.

4.4 Met Constraints

The primary constraint of the project is that the scanning device must be located within the cargo hold of
an aircraft. To meet this constraint our team performed a SWOT analysis to determine the best placement
point within an aircraft. We found through our analysis that the best placement point of our solution is the
cargo door of an aircraft. Our team shared this information with the Connected Digital Systems division
within Boeing which concurred with our findings. Next, we created a graphical model on AutoCAD using
the specifications of the cargo door, ULDs, and scanning devices. We then created a physical prototype of
the AutoCAD model. The prototype had the placement points of the Kinects as well as a makeshift ULD
to perform a scan. Once a scan was complete and matched the physical dimensions to the dimensions
obtained from the scan, our constraint was met.

Professional and Ethical Responsibilities

5.1 Impact of Engineering Solution in Global and Societal Context
As our proposed solution is intended to maximize space in a cargo hold by digitalizing cargo and turning
the unit loading devices into 3D models which will help the loading planner better arrange the cargo into
a plane for best space maximization. However, doing so will not only put more weight on the plane, but it
will also affect the carbon footprint that the cargo plane will leave due to more fuel that will be needed to
pull a heavier plane. Aviation CO2 emissions account for approximately 2.5% of the global average as of
2007. However, nothing is being done to help lower emissions. According to the Boeing World Air Cargo
Forecast (Boeing, 2012), “global air cargo traffic will expand at an average annual rate of 5.2% in the
next 20 years.”

Our solution will provide cargo data and space utilization information from each outstation that the client
is contracted with. As a result, the client will be able to compare which outstations are performing better
than others, if certain outstations perform worse than others, an unintended consequence of the new data
will cause a loss of jobs. This is due to the information not being available previously to measure
performance. The ability to maximize cargo space utilization means more cargo is sent per flight. Certain
outstations send an average of 40% of volume per flight. This means that two flights would be required to
ship goods in one fully maximized flight. As a result, outstations would load fewer planes. The workers
 41

would complete the work in a faster manner which translates to more efficient work. Thus, requiring
fewer labor hours and fewer labor personnel to accomplish the work plan. The outstations themselves
may reduce their workforce with the implementation of our solution. Maximizing cargo means one
optimized plane can ship the amount that two unoptimized planes used to ship. There will be a surge in
airplane purchases due to the innovation. The initial result will be greater control of the market share, a
reduction of competitors in the air cargo industry, and expansion of the company's influence. However,
plane purchases will plateau eventually since one plane can ship what two planes used to.

The increase in plane purchase will have a positive effect on the US economy. Boeing said the current
deal with Qatar Airways will sustain more than 35,000 U.S. jobs and provide the American economy with
an annual estimated economic impact of $2.6 billion throughout the delivery period. "The economic
impact of this sale will reverberate throughout the United States," Raimondo said in a statement after the
signing. "These new freighters will be manufactured and assembled by American workers on American
soil in Everett, Wash." U.S. President Joe Biden hailed the sale as "one of the largest deals that Boeing
aircraft has ever had" in a press briefing before he held bilateral talks with leader Sheik Tamim bin
Hamad al-Thani. [20]

5.2 Met Ethical and Professional Standards

Throughout the design process of our system we held several ethical standards as engineers. Firstly, Act
“for each employer or client as faithful agents or trustees” (NSPE Code of Ethics for Engineers, 2021).
This included relaying all accurate data to Boeing regarding our system whether the solution met all of
the desired objectives or not. As this is a multi-billion dollar contract between Boeing and its client, it was
our responsibility to say exactly how the system met or failed requirements as well as constraints.
Falsifying information at any phase of the design process can lead to large monetary loses and possible
damage to equipment.

Our team will also focus on “hold paramount the safety, health, and welfare of the public”(NSPE Code of
Ethics for Engineers, 2021). In the case of implementing our solution, there will be an increase of cargo
data being transmitted through the Internet of Things. This data can be used to plan disastrous events if
obtained by the wrong hands. Because of this, we worked closely with the Connected Digital Systems
division within Boeing to verfiy that the outputs of our system can be transmitted securely from point to
point.
 42

Works Cited
[1] “Unit Load Devices.” Unit Load Devices (ULDs) | Equipment - Qantas Freight,
https://freight.qantas.com/freight-planning/equipment-uld.html.

[2] “What Is LIDAR? Learn How Lidar Works.” Velodyne Lidar, 28 Apr. 2022,
https://velodynelidar.com/what-is-lidar/.

[3] “Photogrammetry.” Photogrammetry - an Overview | ScienceDirect Topics,

https://www.sciencedirect.com/topics/agricultural-and-biological-
sciences/photogrammetry.

[4] (PDF) Microsoft Kinect Sensor and Its Effect - Researchgate.

https://www.researchgate.net/publication/254058710_Microsoft_Kinect_Sensor_and_Its_E
ffect.

[5] “Sustainability of Digital Formats: Planning for Library of Congress Collections.” STL
(STereoLithography) File Format Family, 12 Sept. 2019,
https://www.loc.gov/preservation/digital/formats/fdd/fdd000504.shtml.

[6] “Boeing 777F.” Aircharter, 29 July 2020, https://www.aircharter.com/boeing-777f/.

[7] “Aero - 777 Freighter.” Aero - 777 Freighter,

https://www.boeing.com/commercial/aeromagazine/articles/qtr_02_09/article_02_1.html#f
ig5.

[8] “Storage and Cargo Handling.” KEYENCE,

https://www.keyence.com/ss/products/auto_id/logistics/role/function-safekeeping.jsp.

[9] “What to Know about Air Cargo Handling.” IATA,

https://www.iata.org/en/publications/newsletters/iata-knowledge-hub/what-to-know-about-
air-cargo-handling/.

[10] Cummins, Nicholas, and Nicholas Cummins (1232 Articles Published) . “The Boeing
777X vs the Airbus A380 - What Plane Wins?” Simple Flying, 11 Nov. 2020,
https://simpleflying.com/boeing-777x-vs-airbus-a380/.

[11] Editor. “Load Planning Software for LCL, Consolidated, and FCL Shipments.” AltexSoft,
AltexSoft, 9 Apr. 2021, https://www.altexsoft.com/blog/load-planning-software/.

[12] Mall, Scott. “Flashback Friday: The History of Air Freight.” FreightWaves, 23 May 2019,
https://www.freightwaves.com/news/flashback-friday-the-history-of-air-freight.

[13] Qiu, Stella, and Jamie Freed. “Air Freight Market Becomes a Bright Spot for Boeing in
China.” Reuters, Thomson Reuters, 27 May 2021, https://www.reuters.com/article/us-
 43

boeing-china-freight/air-freight-market-becomes-a-bright-spot-for-boeing-in-china-
idUSKCN2D80BH.

[14] “Boeing 777X.” Wikipedia, Wikimedia Foundation, 1 May 2022,

https://en.wikipedia.org/wiki/Boeing_777X.

[15] “Digitization.” Wikipedia, Wikimedia Foundation, 16 Apr. 2022,

https://en.wikipedia.org/wiki/Digitization.

[16] “Basics of Photogrammetry – Contents.” Geodetic Systems, Inc,

https://www.geodetic.com/basics-of-photogrammetry/.

[17] Calibration of Multiple Kinect Depth Sensors for Full Surface Model ...
https://www.cse.cuhk.edu.hk/~khwong/www2/conference/2016/IWPR2016/draft_IWPR20
16_cal_mkinects.pdf.

[18] “Ground Operations.” IATA, https://www.iata.org/en/programs/ops-infra/ground-

operations/.

[19] “Digital Cargo.” IATA, https://www.iata.org/en/programs/cargo/e/.

[20] Uria, Daniel. “Qatar Airways Reaches Deal to Purchase Boeing's First 777X Freighters.”
UPI, UPI, 1 Feb. 2022, https://www.upi.com/Top_News/US/2022/01/31/Qatar-Airways-
reaches-deal-purchase-Boeings-first-777X-frieghters/4491643673821/.