Content uploaded by Sakshar Chowdhury
Author content
All content in this area was uploaded by Sakshar Chowdhury on Nov 01, 2023
Content may be subject to copyright.
Study of Robotics & Automation in the Aerospace
Industry
Sakshar Chowdhury
Department of Aeronautical
Engineering
Military Institute of Science and
Technology
Dhaka, Bangladesh
sakshar@ae.mist.ac.bd
Mohammad Mashhunur Rahman
Department of Aeronautical
Engineering
Military Institute of Science and
Technology
Dhaka, Bangladesh
mohammadmashhunurrahman@gmail.c
om
Mohaimen Siddique
Department of Aeronautical
Engineering
Military Institute of Science and
Technology
Dhaka, Bangladesh
mohaimensiddique1@gmail.com
Md. Al Irfan Talukder
Department of Aeronautical
Engineering
Military Institute of Science and
Technology
Dhaka, Bangladesh
accirfan5@gmail.com
Rezaul Karim Nayeem
Department of Aeronautical
Engineering
Military Institute of Science and
Technology
Dhaka, Bangladesh
rknayeem77@gmail.com
Kazi Tauhid Mokbul Hussain
Department of Aeronautical
Engineering
Military Institute of Science and
Technology
Dhaka, Bangladesh
tahiyanhussain7@gmail.com
Abstract— Robotics, and automation are primarily employed
in the aerospace sector for human error minimization, including
increased safety. Workers are no longer needed to conduct
hazardous jobs like picking up heavy objects, performing repetitive
motion duties, or other potentially dangerous tasks. Moreover,
robots also aid in quality control. The overall 32 research papers
have been scrutinized to understand robotics and automation in
aerospace manufacturing, inspection, and assembly. In addition,
Human-Robot interaction and the use of machine learning in the
aerospace industry were overviewed in this research. Cutting-edge
technologies like telemanipulation were used in aerospace
masking to improve industrial robot capabilities. Autonomous
robotic drilling and orbital drilling for machining aerospace
materials were also reviewed. A Significant amount of production
delay, material wastage, and structural inaccuracy will be
avoidable using emerging automation. Optimized automation can
ensure top-notch structural safety, production quality consistency,
and good accuracy than manual methods.
Keywords—Industrial Robot, Automation, Human-Robot
Interaction, Aerospace Inspection, Aerospace Assembly.
I. INTRODUCTION
In the Aerospace Industry, robotics and automation are
interrelated. Due to advancing technologies, the use of
robotics and automation is increasing rapidly in the aviation
industry and outer space. Robotics is a discipline of
technology that teaches machines how to interact with their
surrounding with or without the help of human interaction.
On the other hand, automation does not require any human
interaction to perform mechanical activities in the aerospace
industry. Both automation and robotics help in error
minimization in any activity and lower the workload bested
upon a human being.
Our research emphasizes data-driven technologies, use of
machine learning, non-destructive testing of complex
geometries, automating the riveting of the fastener
components, spacecraft autonomy, telemanipulation-based
aerospace masking skill, development of ultra-lightweight
arms and multi-fingered robotic hands, automated draping
process for dry carbon fiber textiles, introducing friction stir
welding robot, metrology-assisted industrial robots, a work
cell calibration method and many more through use of
automation. Automation can perform activities to develop and
design advanced composite aircraft tail cones (ADVITC) and
follow dry fiber placement (ADFP), tufting, and optical fiber
monitoring as manufacturing functions.
II. ROBOTICS & AUTOMATION IN AEROSPACE
MANUFACTURING
Robots are critical in the aerospace industry, where
automated control enables the production of aircraft, satellites,
and space vehicles. Nowadays, automation enables many
operations such as painting, drilling, riveting, and polishing to
be completed quickly and with minimal effort.
Airbus UK's Robot Capability Tests and Industrial Robot
Positioning System Development for the Aerospace Industry
were presented by Summers [1] in great detail. Four
articulated robots were chosen for the capability test, and their
accuracy and rigidity were assessed using the Robot
Capability Test. The required absolute accuracy could not be
found by enhancing the model, which was kinematic. In
contrast, Positional Data Modification and Pre-filtering
enhanced the accuracy of all test types, and temperature
changes also had a significant impact on positional accuracy.
It will be easier to achieve more accuracy and rigidity by
reducing the working envelope, tool offset, and temperature
variations and using a closed-loop control system to detect
where the robot is about the rest of the work envelope.
Without a closed-loop control system, an industrial robot
cannot achieve adequate position accuracy. Airbus and its
research partners developed the Adaptive Control solution to
give an accurate positioning system capable of controlling the
robot across its whole working volume, monitoring and
reacting to deflection and compensating for external factors
like temperature. A successful test was conducted on an
integrated metrology system for adaptive control. This system
integrated metrology, OLP, a robot, and software interfaces.
Adaptive control enhanced precision in cartesian space while
reducing variation in the final position. Future OLP
development will be dictated by the need for a single or
cooperative robot. For collaboration, a robot will need a
Virtual Robot Controller interface product.
Neumann [2] investigated the key to aerospace
automation. Parallel Kinematic Machines (PKM) are well-
known in the aerospace automation field. All PKM machines
have joint issues which are difficult and complex to create
with backlash-free high stiffness. It was designed to meet the
requirements of the Automotive and Aerospace industries for
production. An electrical robot with high flexibility and a
large work envelope was developed without stability or high
accuracy requirements. Serial Linkage Technology was
applied. Manufacturers had to construct machines with
massive frames and large beds to ensure serial linkage system
precision and rigidity to compensate for Linkage technology.
Nevertheless, it eliminated flexibility. Triceps is a PKM
machine created by Neumann and was the first to properly
integrate robot flexibility and envelope with machine tool
accuracy and stiffness. With its high dynamic capability, a
PKM machine does not distinguish between flat and
compound angle surfaces, making it ideal for all machining
tasks. The PKM machine needed an inordinate quantity of
Inactive Degrees of Freedoms to build a moving box frame.
As a result, other means must absorb applied forces, such as
the middle tube in a Triceps. The new Exechon Concept
addressed all known PKM limitations and achieved all PKM
objectives such as high stiffness, dynamics, and flexibility.
The redesigned Exechon Concept used fewer 1-DOF joints
and actuators with 2-DOS, bending, and linear in one
direction. This design provided a framework to withstand all
torsion and bending forces subjected to the machine.
Kihlman et al. [3] focused on comparing the cutting forces
in orbital drilling to conventional drilling and how this
influenced the results when the same experiments were
performed with an industrial robot. The results demonstrated
that forces were eight times lower in orbital drilling than in
conventional drilling. The stiffness of standard industrial
robots is insufficient to resist deflections due to cutting forces
in drilling. Robotic drilling must be used instead of manual
drilling to automate aircraft assembly. Efforts to use flexible
industrial robots have been tested, but Serial connected arms
reduce stiffness. This is why parallel kinematical machines or
CNC machines have been tried. Orbital drilling works by
bypassing the tool through the material while rotating it
around its axis. The tool's center circles around a static hole
center, reducing force and eliminating numerous issues
associated with traditional drilling. This drilling can drill
cylindrical, conical, and complex-shaped holes. Another
advantage of the force measurements was that the difference
in forces when the tool first exited the material might be
utilized to detect a new layer subsequent in a stack.
De Souza et al. [4] proposed a non-object-agnostic
grasping process. A reconfigurable robot can grasp a software
pipeline that can autonomously produce grasps across a set of
identified items and select the best grasp with task-oriented
capabilities. This pipeline used the Operating System of Robot
and the “Grasp It!” simulator to enhance the application of
Simulated Annealing. The tests were conducted and reported
in the context of an aerospace intralogistics use case. Thirty
iterations were chosen in all cases. Each SA iteration tried
10000 poses for contact point location between active pairs or
fingers. Unmanned bin picking was performed by a human
supervisor using a custom-built grasp viewer software. The
grasp synthesis generated interaction of the initial
circumstances of SA. The stability measure was evaluated to
develop a realistic and non-redundant grasp database. This
software pipeline generated grasp solutions offline and
afterward determined the ideal grasp position by considering
a set of heuristics that strive to create a good grip while
demanding the minimum effort from the robotic arm. The
grasp selection techniques employed were: roll, Euclidean
distance, joint space filter, yaw, and pitch distances.
Caggiano et al. [5] investigated an actual manufacturing
cell's production cycle and performance in the aerospace
sector to improve flexibility and efficiency. Making new
marine gas turbine parts with similar geometry but
significantly heavier than the initial vanes of aircraft engines
produced in the production cell necessitated more flexibility.
The handling robot boosted the cell's flexibility by allowing it
to operate on non-production parts. It would have taken
winches and mechanical manipulators to carry parts manually.
The proposed automation costs 120.000 € for the industrial
robot and 5000.000 € for automatic surface inspection. It was
utilized to examine the cell's behavior before and after the
upgraded automation and production and resource
consumption in both cases. Despite the high cost of automated
equipment, the simulation showed that this investment would
raise the final output volume by 42%. Stochastic variables
were employed to describe machine failures, delayed
incoming batches, and unpredictability in the manual process
cycle time. Adjustments were usually well received by the
manufacturing cell.
Simulation and Digital Twin technologies were reviewed
by Phanden et al. [6] These can run virtual simulation models
though they are not the same. Simulation can be used for
product analysis, manufacture, and design, whereas DT has
much more to accomplish than classical simulation and
unused virtual simulation models become DTs. It gains power
when it receives data that is real-time from its counterpart. If
the design changes in a CAD-based simulation, the simulated
product will behave as it would in the real world. An IoT-
enabled smart thermostat may monitor how well the device is
utilized and make adjustments depending on actual usage
data. DT feeds back data from all phases of the lifecycle of a
product. Scientists can use a digital twin to replicate as many
components of the corporate workflow as they need to
enhance procedures and make business choices. By applying
simulation techniques such as CFD , CAE , FEM , and
simulation of Monte Carlo, various applications-based
simulation techniques are used in the aerospace sector to
simulate the continuous-time history of flights. In
manufacturing, simulation helps simulate design limitations,
human intervention, and other outside interruptions.
Roboticists utilize simulation to test robot control methods.
This study presents recent simulation-based DT models
presented by researchers in aerospace, manufacturing, and
robotics. This review seeks to clarify the conceptual
underpinning of DT for diverse applications.
Dell’ Anno et al. [7] emphasized composite manufacturing
components of the aerospace industry by automated
processes. The activities involved developing and designing
composite advanced aircraft tail cones (ADVITC) using
automated dry fiber placement (ADFP), tufting, and optical
fiber sensor monitoring. The risk of failure was minimized
using tailored fiber placement which fully exploited the load-
bearing capabilities of the fiber. The collaboration of six
robots in automation can increase the potential for the
modularity of systems. Locally reinforcement of critical areas
of the component resulted in a reduction of structural
redundancy by tufting. Monitoring online of resin curing and
infusion is accurately done by optical fiber sensors. This
automation can provide consistent quality manufactured parts
with a reasonable accuracy margin while saving time and extra
cost to manual methods.
Padalko et al. [8] focused on evaluating the efficiency of
integrated automation used in aerospace organizations.
Integrated automation in the Aerospace industry plays a vital
role in solving scientific, environmental, and national security
and defense issues. This scientific challenge was answered
using a strategy based on fuzzy numbers mathematical
equipment, which guarantees the assessment of productivity
and the reliability of automation systems' operation.
Bogue et al. [9] aimed to represent the advancement in
using robots for aerospace applications. This paper provided
extensive usage of automated robots in several unreachable
operations. Additionally, drilling and countersinking holes,
inserting the fasteners, and completing the riveting process are
possible to complete with top-notch accuracy. A robotic arm
of up to eight degrees of freedom is used in coating removal
and composite component production. The usage of 3D-
printed parts increased immensely in modern aerospace
vehicles. Instead of traditional methods, automation can
reduce 55 percent of costs, material wastage by 75 percent,
and also 80 percent of production time. That helped
automation introduce a new dimension in aerospace
component production and maintenance.
Ajay Kumar et al. [10] defined robots as computer-
programmed multi-tasking machines that can efficiently
complete tasks. Most complicated robots are swapped in the
manufacturing sector of aerospace to reduce human errors,
time consumption, and material wastage costs. In the case of
airport security, biometric features and other cargo handling
robots are used. Innovative artificial intelligence in aircraft's
autopilot and communications are already being used as an
element of air traffic control. As the aerospace industry
requires high safety, structural failure due to human errors is
avoidable by using artificial intelligence in manufacturing and
maintenance.
Olfaresultti-Saber Mechanical control systems [11]
consist of lesser controls than the number of configuration
variables termed the underactuated systems. Kinetic
symmetry properties were exploited for the underactuated
systems to reduce complexity. Forced Euler-Lagrange
equations of motion were used to define the system's
dynamics. Class-I underactuated system in zero-gravity
behaves similarly to a mechanical system with first-order
nonholonomic constraints. It has tremendous implications for
space exploration.
Dönmez Özkan et al. [12] came up with mode awareness
of automation driving (AD), aviation (AV), and human-robot
interaction (HRI). Vehicles should not be fully automated, but
the automation should be done to a certain degree, i.e., the
division of tasks to perform by the vehicle on its own and
which the human will. The paper analyzed mode awareness in
three leading functions.
• Mode awareness interfaces: Presentation of
automated mode’s system by any image or verbal
description.
• Methods and techniques: Techniques applied to
encourage mode awareness.
• Gaps: Lack of features of mode awareness
interfaces.
TABLE I. NUMERICAL ANALYSIS RELATED TO AUTOMATION
DRIVING, AVIATION, AND HUMAN-ROBOT INTERACTION
Interfaces
Methods and
techniques
Gaps
Verbal
Graphic
Eye Catching
Multimodal
Details
Additional
Locations
Simplifying
Evaluation
Parameters
Complementary
Models/Algorithm
Insufficiency
Overloaded
Predicting/Monitoring
User Preference
S
u
b
j
O
b
j
AD
9
1
2
1
3
1
3
2
3
7
0
1
7
6
1
7
8
2
8
6
2
2
AV
2
1
2
3
0
8
4
5
1
3
6
1
3
1
5
1
2
1
1
5
1
5
HRI
3
4
5
6
7
6
6
0
4
5
3
4
3
4
0
3
0
Some numerical analysis was revealed in the tabular form
shown above. Existing interfaces were found simple to avoid
complexity without representing full automation features. The
prediction was influential in expanding the range of resolvable
and preventable conflicts. Long-term naturalistic studies were
found to enrich automation capabilities.
Pereira et al. [13] emphasized industrial purpose robots
which were highly flexible with high-speed machine spindles.
The drilling system for the aerospace industry was modified,
fulfilling the requirements of reducing cost, and lead time, and
improving quality. The proposed one is more flexible and can
provide optimized drilling for aerospace aluminum 6061-T6
than a CNC milling machine. With optimized different spindle
speeds, after-condition monitoring minimized the amplitude
of vibrations to the lowest. This drilling system provides better
drilling with improved hole surface roughness and less work
duration for the rapidly growing aerospace industry.
Kizzort [14] focused on the system automation of a
satellite system. According to his research, the Autonomy of
Spacecraft is not the same as the automation system.
However, an automated system is frequently guided by
autonomous features. His research explains a study of the
required language features and takes tradeoffs into account
among proprietary commercial scripting languages,
programming languages, and open-sourced scripting
languages. He has also given an overview of the life cycle
costing vs. the automation level of a satellite system. His work
includes a level scale of six for automation such as Data
Filtering, No automation, fully automated, Cueing, Paging,
and Supervision of ground control systems. He has given the
idea of progressive automation in satellite systems by block
diagram. He has provided how scripting language is chosen in
spacecraft automation. The drawback of the research was the
modeling of the approach called finite-state, which seems
encouraging, especially for the staff minimization required for
manual monitoring and as a level that is intermediate between
the command & telemetry system and an expert system.
However, there are a few amounts of project experience
available.
Angerer et al. [15] proposed a design regarding an end-
effector robot and a complex software solution for performing
automation and the process of draping for dry carbon fiber
textiles. Their research integrates three essential functions
draping, fixation, and gripping. Their design was modular
enough to cope with large and small textiles. The system
design requirements include various dimensions of the
diaphragm, spent, and panel. This paper delivered the concept
of a performing automated end-effector. Moreover, they gave
an idea about secondary and primary lay-up curves on the
diaphragm reference tooling. In addition, their Unified
modeling language (UML) gives an overview of the structure
of the software that they developed for CFKTex. Office
Draping and its Assistant. Their experimental analysis shows
the reachability of the draping, which was required to use a
linear unit for calibrating the location of the tooling separately
for an individual cut. Furthermore, the additional axis had to
be installed into the offline robot programming environment.
Their research successfully automated the most time-delaying
step in Carbon Fiber Reinforced Polymer production. Their
research drawback includes a complete analysis of the quality
level achieved experiment and an analysis of the productive
performance considering draped layers per unit time.
Moreover, a complete analysis of the economy has to be
considered, along with the early and managing costs of an
automated system in the aerospace industry.
Luo et al. [16] introduce the friction stir welding robot in
the aerospace industry. Their research includes a
configuration related to the degree of freedom, system
structure composition, system design of spindle, working
conditions of five compensation principles. Their research
shows that an advanced FSW robot has the capacity of
welding massive and sophisticated curved surfaces in space,
and its process performance is excellent, making it a viable
alternative for accomplishing enhanced metalwork in the
aerospace field. They have also given an overview of the
system structure of the FSW robot. The analysis of the FSW
robot's load and the mixing head force model was one of the
significant portions of the research. In addition, numerical
simulation of the welding process was also elaborated on in
their research. However, no experimental study has been
conducted on the entire project due to the current
circumstances. Current thinking reveals that studying the flow
material in the FSW process will also benefit the FSW robot's
design. Finally, research on the improved FSW robot should
be expanded to include the experimental stage and flow
material in the welding process.
Jayaweera et al. [17] demonstrated the use of metrology-
assisted industrial robots to assemble regional jet fuselage
panels. They have provided a solution for low-cost
manufacturing and less manufacturing lead times. They
concentrated on automated assembly because most of the
work published on robotic installations in the aerospace sector
is based on riveting and drilling applications for aero-structure
sub-assemblies. The conventional approach determines the
actual relationship between the robot TCP (Tool Centre Point)
and the constituent to be assessed about the robot base
coordinate system. It then utilizes that relationship to construct
the appropriate robot programs by modifying a pre-
programmed path. Their system design and development
include end-effector design, non-contact metrology, and a
mathematical toolbox. They had given an overview of cell
construction, cell operation, and error analysis. After cell
testing and evaluation of the system parts arrangement with
mismatched tooling holes were done, this research tested the
automatic construction of typical airplane parts utilizing an
ordinary industrial robot and a laser stripe sensor and studied
to establish that the parts can be assembled automatically.
The influential research done by Leali et al. [18] depicts
the work cell calibration method consisting of four steps to
improve correctness in machining robots in aerospace parts.
Their research mainly showcases the lack of accuracy in
industrial robotics in the case of programming,
manufacturing, machining, and CAD-based simulations. An
approach named robofacturing, which could be efficiently
addressed by specific frameworks design and Process
identification, Resource selection design, and other factors
(PPR) and Product analysis was introduced in their research.
They have given an overview of error compensation in robot
machining, accuracy found in robot machining, and product,
process, and resources. They have described the work cell
calibration method and experimentally validated it. Their
research claim that the peripheral parts' verification and the
calibration of the measuring system of the robot reduce the
errors due to the work cell devices because the robot's tools
calibration and the workpiece alignment reduce the errors due
to the workpiece disparity and the process. Furthermore, their
case study on a robotic cell for high-quality finishing of
aerospace parts demonstrated the feasibility of calibrations as
fully automatic tasks in the robotic process, where correctness
in finishing operations meets the criteria without jeopardizing
the entire system's productivity. Finally, in the actual work
cell, an online tuning carefully checks finishing faults and, as
a result, avoids unnecessary overcorrection.
III. ROBOTICS & AUTOMATION IN AEROSPACE ASSEMBLY
Aerospace assemblers build, install, or repair aircraft parts.
As an aerospace assembler, duties may be limited to a single
element or section of the aircraft. Hand tools or other
equipment to be installed parts or systems in the plane. In this
job, Strict safety and quality standards must adhere to, and
document installations and repairs may be required.
Anscombe et al. [19] presented a collaboration between
OC Robotics and Airbus to develop snake-arm robotics
technology for autonomous inspection and assembly tasks
within aircraft wing boxes. Manual access to rib bays is
difficult due to severe access constraints. Composites are also
a health hazard due to fine dust produced by drilling and the
use of solvents and equipment that makes much noise in small
spaces. Snake-arm robots can provide tools to all parts of the
wing box, replacing manual procedures. With three end
effectors that may be swapped out built by OC Robotics, the
path-following capacity of a snake arm could be maximized
to its full potential. Airbus UK outlined different automation
equipment operation modes as needed for the system to
perform activities within rib bays. If airplanes are constructed
utilizing advanced automation, new tools for maintenance and
repair activities are likely to be necessary. More broadly, the
advancement of snake-arm robotics might allow for critical
design and process changes in the future, potentially saving
the aerospace industry much money. Future development
might use lesser access panels, which would cut down on
maintenance time. The demonstration snake-arm robot's early
results and hopes for future development are presented.
Walton et al. [20] created a model for Robot-Human
Coordination in Aerospace Equipping processes. They made
a significant attempt to automate the aircraft's structural
assembly. Assembling the aircraft's components requires
time-consuming assembly techniques and many complicated
tasks requiring excellent dexterity and judgment from human
operators due to the tight tolerances required for joining parts
and the restricted access required. The flexible metrology-
assisted collaborative assembly, where the pieces are placed
by a robot while an operator fixes or installs them, may be a
potential solution to this challenge. Existing regulation
requires many separation distances and protection systems,
making implementation challenging. A safety control
technique was included that enables the integration of the
robot and operator and manual component repair with the
drives active. To ensure direct cooperation, two principles
were applied. The structures will be used to assess the impact
of examining human factors on the interaction between
humans and robots to facilitate the application of this
methodology: task analysis, human error analysis, mental
workload analysis, user interface analysis, situation awareness
analysis, and user acceptance analysis.
As a Future Automated Aircraft Assembly Demonstrator,
a high-accuracy real-life implementation of Measurement
Assisted Assembly was proposed by Drouot et al., [21] which
is a crucial concept for modernizing aerospace assembly
operations, increasing productivity while lowering costs. This
proposal offered a paradigm change in high-complexity
product assembly by combining robotics with revolutionary
measurement methods. The expected outcomes include
improved component positioning precision and reduced
rectification and rework requirements, which are frequent in
traditional assembly processes, especially not only in aircraft
production. Determining and correcting the position of robotic
manipulators during assembly activities is critical to achieving
these goals. The Future Automated Aircraft Assembly
Demonstrator produced by the University of Nottingham
presents a high-accuracy real-life implementation of
Measurement Assisted Assembly. Experiments showed that
using the production environment outlined below, major
airframe components can be positioned to within 0.1 mm.
Lim et al. [22] reported the progress and assessment of
cognitive interfaces of human-machine and interactions that
facilitate adaptive automation in one too many applications.
New forms of adaptive automation are achievable by
measuring the real-time neurophysiological parameters of the
operator’s intellectual state and providing adaptive decision
support to the user during periods of sustained workload. The
investigation was done to evaluate the potential for a CHMI2
to support multi-UAV (Unmanned Aerial Vehicles)
operations. An unmanned aircraft system (UAS) simulation
environment was developed by allowing participants the role
of a remote piloting an OTM bushfire detection scenario. The
performance varied across participants, with an RMSE of 0.2
to 0.6 for the interred workload. Limitations could be
overcome by introducing different machine learning methods
and additional calibration and tuning stages. This
demonstration of the CHMI2 system paved a new gateway to
future automation in aerospace systems.
Laudate et al. [23] investigated the potential benefits of
human-robot interaction (HRI) for practical aerospace
applications. It was proposed to redesign a workstation that
will assemble a human-made composite fuselage panel and
robots. Under the safety standards ISO 10218 and ISO/TS
15066, the feasibility and functionality of the workplace were
assessed in terms of operating hours and ergonomics. Some
numerical analysis was revealed in the tabular form shown
above.
Significant improvement was observed during assembling
performances with HRI, reducing work duration by 47.6% for
manual operation. OWAS index was evaluated concerning
ergonomics. The results showed that HRI created harsh
working conditions for humans and the risk of task injuries.
Further improvements were possible regarding optimizing the
job sequence and ergonomic assessment quality.
Fig. 1. Human-robot interactions in fuselage assembling.
Weng et al. [24] demonstrated a quick and instinctive-
based telemanipulation method to teach a robot aerospace
masking skills. Their research mainly focuses on how robots
can effectively learn through telemanipulation and improve
their skills based on optimization with sensory feedback. They
have used an offline programming methodology rather than
traditional online programming via teach pendant. Their
proposed method of telemanipulation provides practitioners
with a fast system setup. They can also telemanipulate the
robot with a particular motion, thanks to their method. Their
approval of the technology makes telemanipulation-based
Human-Robot Collaboration a more efficient and natural way
to do robotic aerospace masking jobs. Their research
methodologies include human-robot collaboration in
teaching, fast calibration, adjustable motion retargeting, and
optimization of taping skills. Force regulation and angular
error correction were also used while optimizing. They
experimented with demonstrating the efficiency and efficacy
of the suggested teaching method and a user study to
distinguish the telemanipulation-based mentoring method.
The learning potential of their investigation is limited by the
standard of the pathways taught by human telemanipulation.
Due to the limitations of the robot's sensing capabilities and
the taping tool mechanism, a low-grade affirmation will cause
the robot to fail to learn. In specific experiments, the
misorientation of the taping tool may result in some taping
wrinkles. Also, some gaps in the middle of layers on the
workpiece exterior are out of reach due to the taping tool
geometry, which requests a particular design of the end
effectors for splendid features.
Delving et al. [25] implemented a "one-up" assembly
system that can eliminate activities related to component
removal and make the components assembled at one time. It
can reduce the significant stage of part manufacture and
required space that will benefit low-budget aerospace
suppliers to manufacture quality products at low cost.
Complicated assembly and fabrication of movable trailing
edges (MTE) such as spoilers, flaps, ailerons, and fairings are
possible with consistent efficiency using multiple stages of
automatic robots. To keep consistently in countersink
diameter, hole diameter, and perfect location, pre-
programmed robots can deliver the best performance with
high spindle speed and minimize the duration of production.
Automatic machine scanning will help to avoid unexpected
positioning errors. The primary goal was to introduce a robotic
system that will operate drilling, hole measurement, and
fastener installing functions.
IV. ROBOTICS & AUTOMATION IN OUTER SPACE
Planetary robotics or outer space robotics is widely used
nowadays. Outer space robotics operations are different than
inside earth robotics because their environmental conditions
are quite different. That is why this robotics sector is rising in
the aerospace industry.
An overview of OOS space robotic systems architecture,
verification, and kinematics calibration technologies was
afterward presented by Xilun et al. [26] The number of
organizations and institutes dedicated to space robotics for
OOS missions is expanding rapidly worldwide. Scale and
dexterity were the primary design considerations for space
research robotic systems. There are many different space
manipulators, but they all use probing cones and hook-claw-
capturing devices to move objects about. Unlike industry
manipulator kinematics calibration, on-orbit calibration
procedures for robotic space systems were examined from
four perspectives. Space robots and OOS algorithms were
geared for cooperative objects, making it harder to catch non-
cooperative ones. The end-effector structural design must be
changed to catch friendly and non-cooperative items of
various sizes and shapes. Other challenges that ground-
verification 6 DOF microgravity physical simulation systems
face include testing and verifying systems at the system level
and developing innovative verification technologies to reduce
costs while increasing accuracy. POE-based methods for
building error models may be better than regular ones, but that
is only speculative. Due to the relevance of space
manipulators in OOS, additional tools for measuring absolute
poses should be developed soon.
Hirzinger et al. [27] briefly discussed DLR's (German
Aerospace Center) space robotics expertise through matching
breakthrough initiatives involving systems on the
International Space Station (ISS). They also discussed the
fundamental technologies required to construct a mechatronic
"robonaut" generation with multi-fingered hands and
ultralightweight arms. They have introduced the background
of space robotics ROTEX, the first remotely operated robot in
space. They have worked with the Canadian Space Agency on
several projects, including remote ground control of the Space
Station remote manipulator system, its most famous robot
system. They have innovated lightweight robots along with
modular arms. Their breakthrough in motor technology is
remarkable as DLR's ROBODRIVE was compared to the
finest commonly available motors. They have provided the
Impedance controlled arms controller architecture, which
actively dampens the flexible joint structure's vibrations. Their
contribution to robotics and automation is praiseworthy as
they have built the lightest robot operating in space.
Based on the German Aerospace Center's past, current,
and future efforts, Welder et al. [28] represented highly
automated systems for remote space missions aiming at the
Lunar surface and other celestial objects (DLR). Moon
attracted the scientists because of its position, free of
ionosphere contamination which provided an extra advantage
to observing the Sun and other celestial bodies. It was focused
on developing the autonomy of the robotic explorers for
further complex missions to avoid single-agent mechanical
asset limitations. DLR and CNES jointly developed the
heritage from the Mobile Asteroid Scout (MOSCOT), and in
October 2018, it was deployed on the asteroid Ryugu.
Fig. 2. Artist impression of the MASCOT lander on Ryugu,
photo of the MASCOT lander flight model. dimensions: 29
× 28 × 21 cm, weight:10kg.
V. MACHINE LEARNING IN AEROSPACE
Machine learning theorizes that machines can learn from
data, spot patterns, and make judgments with little or no
human intervention.
Brunton et al. [29] emphasized the data-driven
technologies and use of machine learning in the aerospace
industry. In his paper, the evolution of data-driven aerospace
engineering was explored. An overview of various machine
learning techniques such as supervised, unsupervised,
reinforcement, deep, and physics-informed learning was
given. Diverse Optimization techniques such as Stochastic
algorithms, the role of structure algorithm design, and atomic
operations for non-convex non-smooth functions were
discussed. Moreover, Scalable and robust algorithms such as
randomized linear algebra and robust dimensionality
reduction were also included. Their work also includes
emerging technologies like digital twins, Sensor technology
and IoT, reduced-order modeling, discrepancy modeling,
Uncertainty Quantification, Autonomy, and control. They
focused on aerospace design in Multidisciplinary design
optimization Model-based engineering, i.e., digital twin,
digital thread, design diamond, etc. Their research elaborates
on aerospace manufacturing processes such as advanced
product quality planning, standardization, automation,
assembly, aerospace materials, and composite fabrication.
They gave an overview of aerospace validation procedures
linked with digital twin technology and machine learning
optimization. They have also provided a brief idea about
aerospace services such as autonomous airside support,
elimination of unscheduled maintenance, and loop closure
with design, manufacturing, and testing. They have
contributed to case studies like predictive assembly and
shimming based on previously studied reactive shimming. In
addition, their case study on V-22 osprey and urban air
mobility provides a brief idea about how emerging
technologies like robotics, automation, and machine learning
are taking over the aerospace industry. The possibility of a
data mortgage and paralysis are two of its drawbacks, which
push a change from big data to smart data.
In the aerospace sector, Khawli et al. [30] devised a
method for contentiously automating fastener component
riveting on an aeronautical structure. Their study covers a
variety of round profile extraction methods used to analyze the
geometric primitives of various spheres in tri space for robot-
assisted technology in production processes in the aerospace
sector. To fit the planes on the point cloud and arrange the data
into outliers and inliers, they utilized a Maximum Likelihood
estimate Sample Consensus approach. Furthermore, they
rotated the average direction of the estimated plane using the
Rodrigues formula after downsampling the inliers, and the z-
axis direction is parallel. The Delaunay triangulation was then
applied to the rotated inliers, yielding a successful interval for
tabulating the points placed at the spherical ends of the
perforations from the inliers. After that, they use a clustering
hierarchical technique to divide the categorized spatial
information into three data sets, each corresponding to two
small holes and one big hole. Finally, three spherical profiles
are fitted to the convex hull obtained from the combined data
sets. The method for locating three holes in a tri-point cloud
was commendable. Their technology provides a reference
frame for a robot's experimental riveting procedure with exact
results.
Cassandra et al. [31] focused on ML implementation to
enhance the traditional process of deburring in the industry of
aerospace. They discussed the activities of the industry, which
include machine learning and data communication through
cloud computing. They tried to introduce us to machine
learning to estimate the surface finish, which elaborates on the
analytical approach to ascertain the length of chamfer of the
deburring process by ANFIS ML analysis from sensor
information gathered. The main aim of the research was to
reduce manual workload. They used analytical methods like
an adaptive neuro-fuzzy interference system (ANFIS). In
addition, SVM and FIS to collect the data and analysis to
estimate the surface finish in deburring were also used,
respectively. They discussed the functions of sensors,
displays, ABB robots, and controller equipment. They
mentioned the difference between offline and online analysis
of data. Finally, their approach to solving the traditionally
time-consuming process of aerospace deburring by machine
learning and cloud services was convenient.
VI. ROBOTICS AND AUTOMATION IN AEROSPACE INSPECTION
The work and safety of aerospace bodies like planes,
satellites, missiles, and even spacecraft are examined and
evaluated by aerospace inspectors. It is the job of aerospace
inspectors to comprehend the machine specifications through
drawings and models and inspect such equipment for the
quality of assembly and operation.
Mineo et al. [32] conducted their study by combining
robots and automation in non-destructive testing of
complicated geometries. They created a multi-robot flexible
examination cell with advanced robotics installed on 7-meter
rails and a rotational axis on the outside. The position of the
components put within the robotic working envelope, and
their departure from CAD are initially assessed using their
robot-based photogrammetry approach. Their study focuses
mainly on overcoming NDT's current shortcomings. Their
primary goal was to construct an automated hybrid cell
demonstration to verify and incorporate manufacturing
enablers for future wing designs. It had an ultrasonic non-
destructive assessment and measuring inspection capabilities.
They also gave a quick overview of robotic cell and
instrumentation, outside axes, and basic robot tooling,
revealing that a six-axis articulated robot arm was employed
for floor track or mounting. In addition, a brief introduction of
metrology instruments, NDT equipment, Force Torque Sensor
for correcting real-time, system integration, Robot Cell
Communication Links, and Photogrammetry NDT Inspection
was provided. In addition, for phased array ultrasonic testing,
Offline programming was utilized to build a scanning route.
Photogrammetry images were acquired due to their study for
the geometry assessment of the 3m2 composite wing sample.
VII. OBSERVATION
• In the field of aerospace manufacturing, major
changes were brought to the positional accuracy of
robots, stiffness, high dynamic capabilities, torsion
bearing capability, and flexibility of joint of robots.
Quality manufactured parts can be provided
constantly by automated systems while saving time
and extra cost. An adaptive Control solution was
used to give an accurate positioning which enhanced
precision in the cartesian plane. Best grasp with
task-oriented capabilities was produced using the
Non-object-agnostic grasping process. The final
output value can be raised by a huge margin using
handling robots on nonproduction parts. Receiving
data from real-time counterparts was done through
simulation-based DT models which made real-time
essential changes. Advanced Friction Stir Welding
robot performances were excellent, making it a
viable alternative for accomplishing enhanced
metalwork in the future.
• In aerospace assembly, a lot of huge improvements
were made which resulted in technologies like
snake-arm robotics technology, Robot-Human
Coordination, Cognitive interfaces of human-
machine, Human-robot interaction (HRI),
instinctive-based telemanipulation method, "one-
up" assembly system. Manuel procedures for
assembling in wing box can be replaced by Snake-
arm robots which can provide tools to all parts of the
wing box. By measuring the real-time
neurophysiological parameters of the operator’s
intellectual state, adaptive decisions were supported
to the user during periods of sustained workload
through the cognitive interfaces of the human
machine. Activities related to components removal
and making the components assembled at one time
were done by a "one-up" assembly system. Work
duration was reduced greatly for manual operation
because of Human-robot interaction (HRI). Further
improvements will be possible regarding optimizing
the job sequence and ergonomic assessment quality.
• Outer space robots were different from planetary
robots and expanded rapidly for OOS missions. To
move objects, probing cones and hook-claw
capturing devices were used. OOS algorithms were
geared toward cooperative objects and friendly non-
cooperative items of various sizes and shapes were
caught by the end-effector structural design change.
To reduce cost and increase accuracy, innovative
verification technologies were developed. For
avoiding single-agent mechanical asset limitations,
the autonomy of the robotic explorers was
developed. Lightweight robots with modular arms
remotely operated space robots and impedance-
controlled arms were developed by DLR which
were praiseworthy in the field of space robotics.
• Using the machine learning process, machines were
able to learn from data, spot patterns, and made
judgments with little or no human intervention.
Data-driven technologies were of great use in
aerospace engineering. Overview of various
machine learning techniques such as supervised,
unsupervised, reinforcement, deep, diverse
optimization techniques such as stochastic
algorithms, scalable and robust algorithms such as
randomized linear algebra and robust
dimensionality reduction, emerging technologies
like digital twins, sensor technology and IoT,
reduced-order modeling, discrepancy modeling,
uncertainty quantification, autonomy, and control
were improvements done to machine learning. The
approach to solving the traditionally time-
consuming process of aerospace deburring by
machine learning and cloud computing will be very
convenient for the future.
• Understanding an aerospace body’s features through
drawings and models for assembly, operation, and
safety purposes was the main job of aerospace
inspectors. Robot flexible animation cell with
advanced robotics, assessed by photogrammetry,
was used for NDT of complicated geometries. An
automated hybrid cell, basic robot tooling, NDT
equipment, and robot cell communication links were
demonstrated. The scanning route was illustrated by
offline programming for array ultrasonic testing.
Defect analysis, maximization of the data
acquisition rate, and enabling simultaneous NDT
will be done for better results.
VIII. CONCLUSION
This research involves Robotics and Automation in
Aerospace production, assembly, inspection, machine
learning, and how automated robots are used in outer space
operations. Automation is intended to reduce human labor, but
it should not obliterate human interaction. As a result,
automation should be done so that the distribution of tasks
between humans and robots is optimized. Automating
industrial processes with robotics and other technologies
reduces the risk of human error in manufacturing, especially
when it comes to structural defects. Practical environmental
evaluation can help artificial intelligence in unmanned aircraft
systems increase their safety and efficiency. Machine learning
is critical in aerospace because it enables a variety of
optimization strategies such as stochastic algorithms. In
various aerospace applications, cognitive machine-human
interactions and interfaces enable flexible automation.
However, machine learning has limitations, including the
possibility of data mortgage and paralysis. Human interaction
with Robots can improve systems efficiency by increasing
different mode awareness, evaluation of methods and
techniques, and gap utilization.
IX. AUTHORS CONTRIBUTION
1. Sakshar Chowdhury-Planning and Idea, Abstract,
Introduction, Review of Papers, Conclusion and
Proof Reading.
2. Mohaimen Siddique-Abstract and Review of
Papers.
3. Rezaul Karim Nayeem-Conclusion and Review of
Papers.
4. Mohammad Mashhunur Rahman-Review of Papers,
Observation and Formatting.
5. Md. Al Irfan Talukder- Review of Papers and
Formatting.
6. Kazi Tauhid Mokbul Hussain- Review of Papers
and Observation.
REFERENCES
[1] M. Summers, “Robot capability test and development of
industrial robot positioning system for the aerospace industry,”
2005. doi: 10.4271/2005-01-3336.
[2] K. E. Neumann, “The key to aerospace automation,” 2006. doi:
10.4271/2006-01-3144.
[3] H. Kihlman, I. Eriksson, and M. Ennis, “Robotic orbital drilling
of structures for aerospace applications,” 2002. doi:
10.4271/2002-01-2636.
[4] J. P. Carvalho de Souza et al., “Reconfigurable Grasp Planning
Pipeline with Grasp Synthesis and Selection Applied to Picking
Operations in Aerospace Factories,” Robot Comput Integr Manuf,
vol. 67, 2021, doi: 10.1016/j.rcim.2020.102032.
[5] A. Caggiano, F. Caiazzo, and R. Teti, “Digital factory approach
for flexible and efficient manufacturing systems in the aerospace
industry,” in Procedia CIRP, 2015, vol. 37. doi:
10.1016/j.procir.2015.08.015.
[6] R. K. Phanden, P. Sharma, and A. Dubey, “A review on
simulation in digital twin for aerospace, manufacturing and
robotics,” in Materials Today: Proceedings, 2020, vol. 38. doi:
10.1016/j.matpr.2020.06.446.
[7] G. Dell’Anno et al., “Automated manufacture of 3D reinforced
aerospace composite structures,” International Journal of
Structural Integrity, vol. 3, no. 1, 2012, doi:
10.1108/17579861211209975.
[8] S. N. Padalko, A. A. Ermakov, and F. M. Temmoeva,
“Methodology for evaluating the effectiveness of integrated
automation in aerospace industry,” INCAS Bulletin, vol. 12, no.
Special Issue, 2020, doi: 10.13111/2066-8201.2020.12.S.12.
[9] R. Bogue, “The growing use of robots by the aerospace industry,”
Industrial Robot, vol. 45, no. 6, 2018, doi: 10.1108/IR-08-2018-
0160.
[10] J. K. Ajay Kumar and G. Srinivas, “Recent trends in Robots
Smart Material and its application in Aeronautical and Aerospace
Industries,” in Journal of Physics: Conference Series, 2019, vol.
1172, no. 1. doi: 10.1088/1742-6596/1172/1/012035.
[11] R. Olfati-Saber, “Nonlinear Control of Underactuated
Mechanical Systems with Application to Robotics and Aerospace
Vehicles,” Thesis PhD, 2000.
[12] Y. Dönmez Özkan, A. G. Mirnig, A. Meschtscherjakov, C.
Demir, and M. Tscheligi, “Mode awareness interfaces in
automated vehicles, robotics, and aviation: A literature review,”
2021. doi: 10.1145/3409118.3475125.
[13] B. Pereira, C. A. Griffiths, B. Birch, and A. Rees, “Optimization
of an autonomous robotic drilling system for the machining of
aluminum aerospace alloys,” International Journal of Advanced
Manufacturing Technology, vol. 119, no. 3–4, 2022, doi:
10.1007/s00170-021-08483-4.
[14] B. Kizzort, “Progressive automation in aerospace systems,” in
IEEE Aerospace Conference Proceedings, 2002, vol. 7. doi:
10.1109/AERO.2002.1035310.
[15] A. Angerer, C. Ehinger, A. Hoffmann, W. Reif, and G. Reinhart,
“Design of an automation system for preforming processes in
aerospace industries,” 2011. doi: 10.1109/CASE.2011.6042411.
[16] H. Luo, F. Zhao, S. Guo, C. Yu, G. Liu, and T. Wu, “Mechanical
performance research of friction stir welding robot for aerospace
applications,” Int J Adv Robot Syst, vol. 18, no. 1, 2021, doi:
10.1177/1729881421996543.
[17] N. Jayaweera and P. Webb, “Metrology-assisted robotic
processing of aerospace applications,” Int J Comput Integr
Manuf, vol. 23, no. 3, 2010, doi: 10.1080/09511920903529255.
[18] F. Leali, A. Vergnano, F. Pini, M. Pellicciari, and G. Berselli, “A
workcell calibration method for enhancing accuracy in robot
machining of aerospace parts,” International Journal of
Advanced Manufacturing Technology, vol. 85, no. 1–4, 2016,
doi: 10.1007/s00170-014-6025-y.
[19] R. Anscombe et al., “Snake-arm robots: A new approach to
aircraft assembly,” 2006. doi: 10.4271/2006-01-3141.
[20] M. Walton, P. Webb, and M. Poad, “Applying a concept for
robot-human cooperation to aerospace equipping processes,”
2011. doi: 10.4271/2011-01-2655.
[21] A. Drouot, R. Zhao, L. Irving, D. Sanderson, and S. Ratchev,
“Measurement Assisted Assembly for High Accuracy Aerospace
Manufacturing,” 2018, vol. 51, no. 11. doi:
10.1016/j.ifacol.2018.08.326.
[22] Y. Lim et al., “Adaptive human-robot interactions for multiple
unmanned aerial vehicles,” Robotics, vol. 10, no. 1, 2021, doi:
10.3390/robotics10010012.
[23] E. Laudante, A. Greco, M. Caterino, and M. Fera, “Human-robot
interaction for improving fuselage assembly tasks: A case study,”
Applied Sciences (Switzerland), vol. 10, no. 17, 2020, doi:
10.3390/app10175757.
[24] C. Y. Weng, Q. Yuan, F. Suárez-Ruiz, and I. M. Chen, “A
Telemanipulation-Based Human-Robot Collaboration Method to
Teach Aerospace Masking Skills,” IEEE Trans Industr Inform,
vol. 16, no. 5, 2020, doi: 10.1109/TII.2019.2906063.
[25] R. Devlieg and E. Feikert, “One-up assembly with robots,” 2008.
doi: 10.4271/2008-01-2297.
[26] X. L. Ding, Y. C. Wang, Y. B. Wang, and K. Xu, “A review of
structures, verification, and calibration technologies of space
robotic systems for on-orbit servicing,” Science China
Technological Sciences, vol. 64, no. 3. 2021. doi:
10.1007/s11431-020-1737-4.
[27] G. Hirzinger, N. Sporer, M. Schedl, J. Butterfaß, and M.
Grebenstein, “Robotics and mechatronics in Aerospace,” 2002.
doi: 10.1109/amc.2002.1026885.
[28] A. Wedler et al., “German Aerospace Center’s advanced robotic
technology for future lunar scientific missions: DLR’s Advanced
Robotic Technology,” Philosophical Transactions of the Royal
Society A: Mathematical, Physical and Engineering Sciences,
vol. 379, no. 2188, 2021, doi: 10.1098/rsta.2019.0574.
[29] S. L. Brunton et al., “Data-driven aerospace engineering:
Reframing the industry with machine learning,” AIAA Journal,
vol. 59, no. 8, 2021, doi: 10.2514/1.J060131.
[30] T. al Khawli, M. Anwar, A. Alzaabi, A. Sunda-Meya, and S.
Islam, “Machine Learning for Robot-Assisted Industrial
Automation of Aerospace Applications,” 2019. doi:
10.1109/SMC.2018.00625.
[31] W. Caesarendra, T. Wijaya, B. K. Pappachan, and T.
Tjahjowidodo, “Adaptation to industry 4.0 using machine
learning and cloud computing to improve the conventional
method of deburring in aerospace manufacturing industry,” 2019.
doi: 10.1109/ICTS.2019.8850990.
[32] C. Mineo et al., “Flexible integration of robotics, ultrasonics and
metrology for the inspection of aerospace components,” in AIP
Conference Proceedings, 2017, vol. 1806. doi:
10.1063/1.4974567.
