Students ' Journal Volume 2 I Issue I

Catalysing Transformation: Leveraging Artificial

Intelligence's Potential in the Mining Sector

A Fahmidha Jinan, MS Adhila Maheen, M Fazeeha, KArfaah Meera

Department of Computer Science and Engineering, Arunachala College ofEngineering for Women
Manavilai, Vel/ichanthai

ABSTRACT
The value chain in the mining industry is slowly evolving to improve the management of existing assets
and increase production. While the industry has gone through ups and downs, there has been no significant
change in the nature of mining. In order for the mining industry to achieve efficiency and effectiveness,
the way it works needs to be re-evaluated. Considering the key challenges that must be overcome for
the success of Al transformation in hazardous mines, the significance of Al revolution in the mining
indus try is explained, focusing on improving safety, productivity, environmental safety and reducing
mining accidents, environment. The use of technology shows that with the development of technology,
the process of my products has improved and the rate of people making mistakes has decreased. This
article provides a unique history of technological and digital developments achieved in specific areas
of the mining industry, which, intentionally or unintentionally, correspond to the products of the Fourth
World Economy. These developments also reveal the gap between what can be done and how to ensure the
health of food. Overall, this article describes the many advancements made over the years in the mining
industry as technology development helps reduce the risks and hazards in work. Al is revolutionizing the
mining industry by enhancing operational efficiency, safety, and decision-making.. In the absence of past
data, Infosys Nia, an Al system, employs machine learning to produce precise predictions. It improves
operational effectiveness, helps with resource allocation decisions, and adds value to a region. Robotic
devices that operate in physically challenging conditions include drones and autonomous trucks.
Objective:
The objective of this work to utilise artificial inte lligence (Al) technology in the mining industry to
improve operational effectiveness, safety, and sustainability while maximizing resource utilization and
reducing environmental impact
Steps:
The steps that are being carried out are:
Obtaining and Organizing Data, Assemble pertinent information from sensors, mining operations, and
other sources, Model creation and instruction, Create Al models specifically for mining operations like
resource optimization or predictive maintenance, Implementation & Integration, Integrate Al into the
mac hinery and processes used in mining. Deploy Al models to real-world settings so they can make
judgments instantly and Monitoring and Upkeep:
Keep an eye out for anomalies and system perfonnance in Al syste ms. Update and maintain Al models
frequently to ensure their dependability and capacity to adjust to changing circumstances. Regularly
monitor the quality and integrity of the data used for Al modals.
Keywords: Autonomous driving-robot, 3D tunnel mapping, Location estimation, wireless, Robot Sensors.

Ill T he lnstilulion of Engineers (India)

 Volume 2 I Issue I Students' Journal

INTRODUCTION Robotics Development in Mine Industry

In this ever modernizing technological world, with Applications
the use of artificial intelligence (Al), the mining According to Coiffet ( 1983), robotics emerged
industry is going through a radical change. One of as a result of the desire to design a machine that
the most noticeable trends is the rise of autonomous could perform all of the tasks that a person could
mining vehicles, where haul trucks, drills, and not or did not want to perform themselves but
loaders powered by Al work independently to nevertheless needed or wanted to accomplish,
improve efficiency and safety [I). By foreseeing without endangering the authority of the human [6].
equipment failures, predictive maintenance, Figure 2 shows a schematic of a typical industrial
made possible by Al, is cutting downtime and robot.
maintenance costs. The effectiveness of Al in The development of robotics in the mmmg
environmental monitoring is essential for ensuring industry has witnessed significance advancements
legal compliance and reducing the environmental throughout the years. In the early days, mining
impact of mining operations. For instance, the operations were often dangerous and labour
development of devices to increase underground intensive with workers exposed to hazardous
mine safety or automatically transport materials condition. However, as technology progressed so
utilized the autonomous driving technology did the role of robotics in mining. One notable
[2][3]. Al algorithms process geological data, innovation is the use of autonomous mining
mineral exploration is becoming more precise vehicles which have improved safety by reducing
and economical, hastening the discovery of the need for human operators to work in perilous
potential mineral deposits. Together, these trends environments. These vehicles have advanced
highlight Al's critical role in streamlining mining sensors and navigation systems that allow them to
operations, boosting safety, and encouraging manoeuvre through difficult terrain and complete
environmentally friendly business practices. In tasks like drilling , transporting and exploration
dangerous occupations like mining, using advanc,ed additionally drones, and unmanned vehicles have
technology is the most efficient way to increase been employed for surveying and monitoring [7].
output and guarantee safe working conditions (4). Allowing for more significant and accurate data
The incorporation of artificial intelligence (Al) collection. In underground mining, remote control
has caused a radical change in the mining sector. robots have become crucial for tasks like inspection,
The proliferation of autonomous mining vehicles, maintenance and even rescue operations. These
where Al-powered haul trucks, drills, and loaders robots can access tight spaces and areas with poor
operate independently to improve efficiency air quality, mitigating risks to human workers.
and safety, is one of the most noticeable trends. Moreover, the integration of artificial intelligence
By foreseeing equipment failures, Al-enabled and machine learning has enhanced the predictive
predictive maintenance is lowering downtime and maintenance of mining tools, reducing downtime
maintenance costs. For reguJatory compliance and improving overall productivity. As the mining
and reducing the environmental impact of mining industry continue to evolve it is likely robots will
operations [5], Al's strength in environmental play and increasingly vital role in enhancing safety,
monitoring is essential. Geological data is processed efficiency and sustainability in this sector.
by Al algorithms, accelerating the discovery Many publications and many developers have said
of potential mineral deposits while improving that in over time , automatic robots can perform
the accuracy and cost-effectiveness of mineral some underground mining operations faster than
exploration. Together, these trends highlight Al's humans, can be constant and develop more reliably
crucial contribution to the mining industry's efforts than human beings, and can work in dangerous areas
to improve safety, streamline operations, and [8]. They are also cheaper than using human workers
promote sustainable practices. for repetitive chores; Frankly, there are some jobs

T he lns titutiou or E ngi neers (India)

Students' Journal Volume 2 I Issue I

that can only be done by humans. Unfortunately, as to monitor the operation of the coal face and to
Strong ( 1980) points out, this expects quite a lot of some extent increase the use of coal transportation.
training and demanded working [9].
Although progress has been made in the past, little
The surroundings for a fixed or changing location progress has been made in preserving facial function.
are determined by the environment a robot is put in These initial shortcomings can be linked to three
and where it functions [IO]. The robot is exposed to things: (I) the way survey data was presented, (2)
hazards that must be avoided and interesting objects the limitations of the instruments at the time, and
that require action. As a result, communication (3) the absence of communications technology for
involving the automatic robot and the surroundings. gathering, transmitting, processing, and displaying
For instance, a robot could operate a underground device information [ 13]. This article only addresses
penetrating tool that could pierce rock at. It must the employment of robots in the field. Automatic
hold on to the sidewalls of its hole while drilling a systems turned into robotic systems. According to
new space at angles greater than 60 degrees to the Figure I, industrial applications have two stages:
vertical before entering it. Robots that are flexible and resilient constitute the
ln order to reach places where it is impractical initial stage; Robots are programmed to perform
to drill underground straight downwards, it is many repetitive tasks. These robots constitute
necessary to drill around bends, curves, sideways, most of the robots used today. Interactive robots
or even upwards. In order to avoid drill breakdown that communicate with their physical environment
due to weight, the autonomous robot being tested represent the beginning of the second evolutionary
for rock drill needs to keep an eye on the drilling phase.
and rock requirements. This year, an all-purpose Artificial intelligence (Al) has completely changed
2-boom drill rig that can work unattended in a mine the mining business, particularly in terms of
direction was created. mining and geographic design (14]. However, the
SURVEY mining sector continues to be one of the industries
with a limited variety of Al and computerized
Recently, Canadian research and industry advancements (15]. The applications of Al have
organizations such as the National Research changed the sciences, advances, and execution
Council (NRC), the Canadian organisation for of the mining industry. Shrewd mines, whicn
Mining and Energy Technologies (CANMET), use progressed advances based on Al, have been
and government agencies (Bartel, 1983) have presented and broadly received around the world.
been using robots and artificial intelligence (Al) to A comprehensive overview of Al applications
address mining problems [ 11 ]. One of the best ways in mining and geographical designing uncovers
to increase production and ensure safe operation in their effect on different perspectives such as
a hazardous industry like mining is to use advanced rock mechanics, mining strategy choice, mining
technology. Although it is not a new concept, the nardware, drilling-blasting, incline steadiness,
idea of fully automating the extraction process is natural issues, and important geographical
rejected since it is not feasible. designing.
All automated mines have been deactivated. In this paper, we use the term ..robotics" in both
possibly ten years. With the installation of definitions and refer to robot-related procedures.
conveyors and cutters at the coal surface years Robots can be utilized in three main mining-related
ago, coal mine automation had its start. The use activities: exploration, production, and helping
of underground mining machines, especially people with challenging or hazardous work. Robotic
those associated with coal surfaces, accelerated exploration operations make it possible to conduct
in the mid-1950s and was completed by the mid- work in dangerous situations, such as those that are
l 970s[ 12]. With this technology, there is an effort underwater, underground, radioactive, or nave high
temperatures and humidity.

II The Lnstitution of Engin eers (India)

 Volume 2 I Issue I Studen ts' Jo urnal

This paper gives a comprehensive set of Al equipment like loaders, drilJs, and haul trucks.
applications in mining and topographical planning. These vehicles use artificial intelJigence (Al) for
lt covers subjects such as rock mechani cs, assurance navigati on, collision avoidance, and effective
of mining methodo logies, mining equipment, material handling , improving both productivity and
drilling- blasting, slant robustness, characteristic safety [ 16], as shown in Fig3.
issues and related topograp hical planning.
ii. Predictive Maintenance: Researchers and
professionals in the business a.re currently looking

_
-- ..
towards Al-powered predictiv e maintenance
solutions. These systems use sensor data from
mining equipment to assess equipme nt faults in
order to forecast and prevent them, eventually
cutting downtime and maintenance expenses.
iii. Mineral Exploration: Geological data analysis
is used with A l and machine learning techniques
for more precise mineral prospecting. This entails
Figure I : A locating possible mineral resources as welt as
timeline or robotics technology
developm ent. streamli ning the drilling and sample procedures.
iv. Operational Optimiz.ation: Drilling, blasting,

r
-.. ,
I
material transportation, and ore processing are
just a few of the mining activities that are being
optimized using Al-driven solutions. Increased

- --- - efficienc y and cost savings are the results of these

- 1=·
optimizations (17].
v. Environmental Monitoring: The research
emphasizes Al 's function in mining sites• ongoing

- environmental management and monitoring. Al

systems aid in the detection and mitigation of
environmental concern s, maintaining regulatory
compliance and minimizing the environmental
Fig 2 A schematic illustration of industrial robot impact of the industry.
Figure 2 shows a schematic illustration of industrial vi. Safety and Risk Management: Research focuses
robot on artificial intelligence (Al)-powered safety
systems that track and examine data from many
This paper first discusses robotics ' uses in mining
sources, including wearable and sensors, to detect
before considering the potential use of tele
and stop accidents and harmful situations.
operatio n in the industry partially for mining, but
primarily for the upkeep of automated equipment
in mines. The issues with identifying the amount of
automation required for the integration of cutting-
edge technological development into mining are
then covered , along with alternative solutions.
PROPOSED METHODOLOGY
i. Autonomous mining vehicles: Numerous studies
investig ate the use of Al for autonomous mining Figure 3: Autonomous mining vehicle

The Insti tution or Engineer s (Ind ia)

 Volume 2 I Issue I
Srud ents ' Jour nal

on mod el's
ng supp ly exce eds the thres hold, the com paris
vii. Supp ly Chai n Man agem ent: For mini ating that the
Al is bein g appli ed. It supp orts valu e state men t is outp ut, indic
chain mana geme nt, If the scor e is lowe r
inven tory mana geme nt, logis tics com paris on mod el is corre ct.
cost-effec tive , lhe valu e: expr essio n recei ved
astin g whil e than the thres hold
optimization, and dem and forec IM U sens or is used by the func tion by
abili ty. from the
ensu ring time ly material avail lar patte rn.
displ ayin g the nega tive of the simi
initia tives 0.5 seco nd
viii. Ener gy Efficiency: Al-driven Mea sured head ing valu es are adde d in
ener gy use have been inspi red by e the curre nt head ing. The
to optim ize incre men ts to dete rmin
gy use. Al usin g the
the mini ng indu stry' s exce ssive ener auto mate d robo t's posi tion is estim ated
y cont rol mach inery analo gue or
algo rithm s are used to effectivel valu e expr essio n obta ined from the
ump tion and rmin ed
and proc esses , lowe ring ener gy cons IMU sens or as well as its mov emen t is dete
ct tions (I), (2),
envi ronm ental impa by the enco ding detec tor [ 19]. Equa
tion: The as a repre senta tion of the auto mate d
ix. Data Anal ytics and Visualiza and (3) serv e
significance of A l-dri ven data analy tics and robo t's appr oxim ate locat ion balance.
ed in the litera ture.
visualization tools is emp hasiz
mass ive datas ets prod uced by z(,.. ,) = z(lt) + d(lt) ·C011{P(tt)) ·C011(o(tt)) ( I )
Thes e tools anal yse
offer usefu l infor mati on for
= r (lt) + d(ft) · COIi (P(t.)) · mn (a (ft)) ( 2
min.i ng activ ities and )
esses . i,(tt+i)
maki ng decis ions and streamJjning proc
inve stiga ting + d(t•) ·m(P (t•))
x. Robotic Mining: Researche rs are z(t...,i) = z(ft) (3)
t be comb ined for work in
how Al and robotics migh
the auto mate d
dang erou s or dista nt mini ng sites . For expl orati on, Of these , d(t.k) repre sents the gap of
tic syste ms are tk., (t.k) repre sents the path
main tenan ce, and inspe ction s, robo robo t at the prec ise time
(t.k) signi fies the step angl e of the
being deve lope d [ 18]. of the robo t and
digit al twin robo t.
xi. Digital Twins: The use of
ing. Real - Whe n start ing the task, the remo te cont rol wire lessly
techn olog y in mini ng oper ation s is grow
izati on of send s a sign al to the auto nom ous robo t. It then mov es
time modeUi.ng, mon itorin g, and optim
ible by these throu gh the mine at high spee d in the cent re of the
comp licat ed proc esses are made poss
prod uctiv ity and hole using horiz onta l ljdar sens ors and auto nom ous
virtual repli cas, whic h impr ove
com parin g
decis ion-m akin g. drivi ng algo rithm s, whil e anal ysin g and
. Estim ated posit ion
rol: Rem ote imag e chan ges of the wall hole
xii. Rem ote Oper ation and Cont rded by the verti cal
equi pme nt cont rol and tunn ellin g data reco
operation and mana geme nt of mini ng ated posi tion cont rol and
unic ation shap ed lidar sens ors, estim
is mad e possible by AI and comm urem ents of poin t data sens ors
irem ent for verti cal lidar meas
technologies, elim inati ng the requ to crea te 3D map s of unde rgrou nd
regions and are also used
hum an prese nce in haza rdou s lated and
impr ovin g over all safety. mine s [20]. All sens or data is calcu
r the eval uatio n,
JD reco rded ever y 0.5 seconds. Late
a Loc1 1tion a nd Map ping a ated posi tion is estim ated
E stim ating the prec ision of the estim
Tunn el the auto mati c robo t's posit ion.
acco rding to
patte rn and
The diag ram show s the matc hing Within the mini ng indu stry, inde pend
ent drivi ng
the sens ors to deter mine und mine s.
sequ ence data used by robo ts are utiliz ed to expl ore unde rgro
the auto nom ous robo t. Hori zont al appr oxim ation
the pos ition of ts requ ire exac t area
data (SK, SK + I), unco nven tiona lly Thes e robo el plan ning to func tion
shap ed lidar 3D unde rgrou nd tunn
to matc h and
calcu lated twice (K, K + I), are used effec tivel y. A rece nt s tudy distr ibute d with in
valu e repre sents the nce and
the patte rns, and the slope national jour nal of Mini ng Scie
ine's nam e and scor e valu e the Inter visio n-ba sed
autom atic mach Inno vatio n, prop oses a mach ine
the actua l outp ut. Whe n the scor~
representing

II The lo5ti tutio o of Engi neers (Indi a)

 Volume 2 I Issue I Students· Journal

design coordinating procedure for independent point cloud layout of the entire underground mine
driving robot area estimation and Mining areas flow. When subterranean arranging is necessary,
underground with 3D tunnel mapping. By using a a neighbouring LiDAR check is then given. To
Light Location and Extending (LiDAR) sensor to eliminate point highlights, quick point highlight
continuously scan the tunnel divider within the flat histogram (FPFH) descriptors are computed for the
course, the robot may execute design coordination global and local point clouds. Using subjective test
by identifying the geometry of the divider [21]. A agreement (RANSAC) and iterative closest point
sensor is connected to the inertial estimation unit's (ICP) selection. the arrangement ranges amongst
sensor in order to determine the automatic robot the global as well as the local point clouds are
position, and an encoder is connected to the sensor examined and adjusted. The products zone on the
to determine the robot's heading. The vertical around the world using this system is measured
LiDAR sensor filters the vertical direction of utilizing the LIDAR sensor [22].
underground tunnelling mine in expansion, and the
Tunnel Mapping in 3D
layered data creates a 3D contour of the mine. This
is done while the robot is moving. Past ponders The LMS-111 type LiDAR radar used in this
were primarily focused on how to implement the research has a range of 270° and data points are
suggested plan since the x and y makes standard measured in 0.25° steps. Since the latency of the
ultimate mistake. lidar sensor is 0.5 seconds, horizontal/vertical lidar
collects 2162 data points in 1 second, respectively.
Flowchart
The coordinates of points in LiDAR that are
Figure 4 shows florchart representing the estimation distinguished by an angle Oas well as the distance
of the position of an automated robot. (D) can be determined using Equations (4) and (5)
[23].

Xpoint Dpoint • cos 8point

(4)
Ypoint == Dpoint • sin Bpoint
(5)
To provide details of the aerial measurement cloud
and create a 3D image, the location and behaviour
of the rover must be collected along with point data.
The roll, pitch and yaw of the robot first determine
its 3D behaviour in terms of Euler angles. The
rotation matrix of the equation. (6), (7) and (8) are
used to represent these and represent the rotational
movement of the automatic robot on the x, y and z
axes correspondingly.

[
l O O
Ri (11111) • 0 a.o - lblo
0 •a a.a
l
Fig. 4. Flowchart representing the estimation or the
(6)
position or an automated robot

Within the mining business, 3D zone surveying

may be a crucial responsibility. Point cloud (7)
registration and flexible LiDAR mapping are two a.,. -•,. ol
methods used for this reason. This plan includes
using flexible LiDAR to create an early global
R1(1n) •
[•-,0 a.70 O1
(8)

The Institution of Engineer s ( lmlia)

II
Students' Journal Volume 2 ( Issue I

Using the pos111on, pose and point information decision-making, extract value from data, and
received from the autonomo us robot, the increase safety I. One of the areas where artificial
coordinate s of the tunnel map are determine d as intelligenc e is making a significan t impact is 3D
shown in Equation ( I). (9). The coordinates of the modelling. Robots can also be used in environme nts
indicated data are xmap, ymap, and zmap, and the where soil is heavy. Tunnel mapping research using
coordinate s of the driver are xrobot, yrobot. and self-drivin g robots [25] and 3D LiDAR sensors.
zrobot. Equation matrix. Equations (6), (7) and (8) detailed design2. This will help engineers make
are combined into a matrix and calculated in z-y-x more informed decisions and increase the efficiency
to make the text consistent with the autonomo us of tunnel construction2. Artificial intelligenc e
robot's posture in the cloud data to draw the hole. innovations in the mining industry have changed
Additionally, xLiDAR, yLiDAR, and zLiDAR the fundamentals of the mining industry, leading to
represent the distance equal to the robot's position increased efficiency and cost savings. The use of
and the object's position, and point and dot are artificial inteUigence in the mining sector will save
the literal coordinat es of the point seen from the raw material production from $290 billion to $390
vertical LiDAR radar [24]. billion annually by 2035 [26]. These advances are
The mining industry is advancing artificial changing the mining industry and paving the way
intelligence (Al) to improve processes, enhance for a more profitable and profitable future.

co,aco,/3 co,ann/3,in,y - nnacor y co,annf ko•'Y - nncuin, y

= ,inaco•/ J nna,in{J ,in-y + co,acor y ,,nannf ko,,y - co,a.n n,y
[
-nn/J co,fJnn, y co,fko,, y

][:=] + [:::] + [ : : ]

S1t •
[
x..,.,
x.-m
.

x.,1""1 Y.-111111
z.1.,
Z.-m

z.,i,.,,
l meter in 63 seconds. The place where the robot
must tum at a certain angle compared to other
(9)
car can verify the safety of passing through a depth

places is as seen in the picture. As shown in Figure

(10) 9b, the consensus model matching does not work or
In Eq. ( I 0), the form (SK) can be used to express its accuracy decreases .
the point data, which includes the position, posture,
In this study, the angle in the yaw direction was
and distance informatio n from LiDAR at time K.
measured using a 2D LiDAR sensor, whereas the
R ESU LT angle in the roll and voice commands were solely
As Shown in fig 5, a robot who is free to walk on measured using an IMU sensor. Integrated models
the ground and perfonns estimation a nd 3D space in yaw, roll and motion directions can improve the
operations. Each section features camera screens precision of the automatic robot's 3D behaviour
and horizonta l lidar capture images of the tunnel. size as well as guarantee the exactness of deep-
As the robot moves, calculate its attitude and quality 3D form. Due to this reason, there is a need
position using direction and distance information to use 3D LiDAR or vision cameras to research our
captured by the sensor in real time. A driveriess behaviour al treatments . Moreover, the test field is
a 25 m long vertical shaft without any interferen ce

The l nslilulion of Engineers (lndiu)

 Volume 2 I Issue I Students' Journal

in the direction of movement. However, in a real To show variation ooJy, variables were calculated
underground, there are obstacles such as personnel, as the mean of values. They show the heading
access restrictions and intersections. deviation between the previous point and the
The model comparison results of the automatic current point. The overall average is close to 2.5°,
driving robot experiment are shown in Figure 6. which means there is a 2.5° rotation in the(+) or
During the movement change of the robot, model () direction every 0.5 seconds. On the other hand,
comparison and accurate position estimation are when the model comparison is not complete and
made every 0.5 seconds, a total of 130 times. the accuracy is 70% or higher, the points are
Comparisons of models are made every time; approximately l.92° and 3.05° respectively. The
however, when the initial standard benchmark research discovered that the design identical
score is 70%, the success and failure rates are precision was good when the automatic robot turned
64.8% and 35.2%, respectively. an angle smaller than the previous angle, while the
accurateness was lower when the automatic robot
turned a larger angle. Class changes sometimes
vary larger than 10°, confirming that the model
comparison accuracy is low.
.

;. - -
. . -
- ·. .
-.., (
Some bends in underground mines require the
automatic robot to tum at a greater tilt. Once the
autonomous robot reaches a curve the price chart
changes significantly and remains that way until
it reaches the next curve. Although the robot
maintains a high value when entering the curved
section, it can be concluded that the comparative
efficiency of the model is higher at the curve point
~ than at the curve point entry.
i

DISCUSS ION
In this study, lMU sensors were installed
corresponding to the model comparison result
to estimate the driver's stationary position.
Additionally, a 3D map was created using vertical
Figure 5: shows a view of an automatic driving robot,
sections of the soil and the position of the robot.
a webcam image, and tunnel waU measurements
made by a LiDAR sensor These measurements were made using 2D lidar.
Estimate the performance of the estimated position
and 3D route map proposed in this paper, a
mathematical judgement of the machine's estimated
position, the 30 tunnel, and the a.ctual results
should be made. The accuracy of the intended
benchmarking model based on job prediction have
been extensively evaluated.
The Figure 7 displays a 3D map made from
.. vertically shaped Li DAR data while the self-driving
autonomous robot travels under the mining tunnel.
Across a magnitude of 25 meters, the autonomous
Figure 6 : shows the result of pattern matching while
robot moves precisely 6 meters in the y-axis region
the automatic robot drove and the 70% score line

The I nstitution of Engineers (Ind ia)

Students ' Journal Volume 2 I Issue I

and 25 meters in the x-axis region. The underground many places that are quite unsafe and inaccessible
tunnelling mine's Z-axis height is measured every to humans. Exploration can be made safer in these
0.5 seconds while moving slowly. areas by using autonomous robots. Additionally,
using low-power vehicles to transport materials in
As well as, Small autonomous robots were used
underground mines can increase mine production.
in the field experiments of this study, but there
may be some areas in deep mines where the road Precise accuracy and 30 mapping are very
is not good. In order for the autonomous robot to important for the use of autonomous robots in
operate in safety in the environment, it is essential these areas. Future research involving position
to develop driving by following the autonomous approximation and 30 planning of robot control in
driving algorithm obtained in this research. underground tunnelling mines should build on the
findings of this research.
CO NCL USI ON
This work uses model matching technology based
on machine vision to realize task prediction and 30
expression of the driver's driving.
I. By measuring the robot's head angle using
a horizontal LiDAR sensor, the proposed
position estimation identifies and matches the
geometry of the tunnel wall. The [MlJ sensor
is used to determine the head when the pattern
matching accuracy is below par.0. Directional
and encoding sensors (which also measure
range and IMU) are used to determine the
robot's model and position; The prediction
is combined with data from vertical LiDAR
sensors to build a 3D underground tunnel. When
Fig.7 3D pattern or a underground mine created by the exactness of the plan was compared to the
driving robot constructed on matching model previous location estimate, it was found that
Dangerous problems in underground mines include the accuracy of the plan increased by 0.07 m.
Additionally, autonomous robot-based tunnel
toxic gases and explosives. Therefore, there are
mapping showed an RMSE of0.05 m2.

Table I . Experiment findings for the mapping and surveying or the underground tunnel segment done by the
automatic self-driving robot
SI NO. Area of section Area or section Absolute error Root mean
created by created by (m2) square error
surveying (m2) autonomous (m2)
driving robot
(m2)
I 6.62 6.69 0.07 0.05
2 6.73 6.66 0.07
3 6.73 6.62 0.1 I
4 5.97 5.96 0.01
5 5.44 5.44 0.00

II T he Institution of Engineer s (Indi a)

 Volume 2 I Issue I Students' Journal

II. Early studies using 3D maps of underground REFERENCE

mines created by nnmaooed robotic vehicles l. Z. Ren, L. Wang, L. Bi , " Robust GICP-based
often lead to errors depending on the type 3D LiDAR SLAM for underground mining
of sensors (such as IMUs and encoders) environment", Sensors, vol. 19 ( 13), pp. 29 15, 2019.
used. When used as a comparison to existing Larsson J, Broxvall M and Saffiotti A,"A navigation
2.
measurements or maps, preliminary research system for automated loaders in underground mines",
should be conducted on studies that use Proceedings of the 5th International Conference on
distance measurements (e.g. lidar sensors) to Field and Service Robotics. Port Douglas, Australia:
measure location. Due to the features of the Springer; 2006. pp.129-40, 2006.
mining soil, it is still challenging to ascertain 3. J. MarsbaU, T. Barfoot and J. Larsson, "Autonomous
the essentials. One more constraint is the extra underground tramming for center-articulated
effort required to determine the shape of each vehicles", J F Robot, vol. 25 (6- 7). pp. 400-421 ,
wall, making it difficult to identify details. Job 2008.
forecasting services have been very successful 4. M.A. de Miguel, F. Garcia and J.M. Armingol,
and can be used in underground mining. "Improved LiDAR probabilistic localization for
autonomous vehicles using GNSS'', Sensors, vol.
111. lo this study, a plane 2D LiDAR sensor was used 20 (11), pp. 3 145, 2020.
to degree the angle in the yaw direction, and
5. Roberto Noriega and Yashar Pourrahimian, "A
only the IMU sensor was cast-off to calculate
systematic review of artificial intelligence and
the angle in the roll and voice commands. The data-driven approaches in strategic open-pit mine
integrated yaw, roll and motion model increases planning, Resources Policy", vol. 77, 2022.
the precision of the automatic robot's 3D
6. Coiffet, P.• and M., Chirouze, "An Introduction to
behaviour quantity and ensures the precision of Robot Technology", Hermes Publishing (France)
the 3D image in depth quality. For this reason, 1982. Strong, D.R. , Ro botics and Mining, 82nd
3D Li.DAR or vision cameras need to be used AGM, Canadian Institute of Mining and Metallurgy,
to conduct research on the treatment of our Toronto, Canada April 20-23, 1980.
behaviour's. Moreover, the testing area is a 25 7. Janakiraman, C.R., Das and B.M. (1988),
m long vertical shaft without any interference Application of Artificial Intelligence and Robotics
in the direction of movement. However, in a to Mining'' , Radharamanan, R. (eds) Robotics
real underground, there are obstacles such as and Factories of the Future '87. Springer, Berlin,
personnel, access restrictions and intersections. Heidelberg. 1988.
8. D. Banerji, R. Ray, J. Basu and l. Basak,
IV. Small robots are used in the field experiments " Autonomous navigation by robust scan matching
of the study, but they may have disadvantages technique", Int J Innov Tech Creal Eng, 2 (2012),
in deep mines. In this research. so that the robot pp. 7-13.
programme may be safe in the environment,
9. Strong, D.R., Robotics and Mining, 82nd AGM,
it is crucial to advance driving execution Canadianlnstituteo£Minin.gandMetaUurgy,Toronto,
using utilizing the automatic driving method Canada April 20-23,1980.
developed on the road.
10. H. Kim and Y. Cho~ "Comparison of three location
V. Accurate estimation and 3D planning are estimation methods of an autonomous driving robot
for underground mines" , Appl Sci, vol. IO ( 14), pp.
particularly important for the use of unmanned
4831, 2020.
robots in these projects. Future research
involving po:sitioo estimation and 3D plotting 11. Bartel, E.W., " Industrial robotics potential
for robotic control in tunnelling mines should applications to mining, Mining Technology and
Policy issues", American Mining Congress 1983
develop from this work.
Mining Convention, September 2-14, 1983, San

The lnstiCution or E nginee rs (India)

Students' Journal Volume 2 I Issue I

Francisco, Cal.,16pp, 1983. tunnel mapping in underground mines using pattern

matched LiDAR sequential images", International
12. L. Bamewold and B.G. LoUennoscr, "Identification
Journal of Mining Science and Technology, vol.
of digital technologies and digitalisation trends in
31(5), pp. 779-788, 2021.
the mining industry", Int J Min Sci Tech, vol. 30 (6),
pp. 747-757, 2020. 20. Stahl T, Wischnewski A, Betz J and Lienkamp M.,
"ROS based localization of a race vehicle at high-
13. Larsson J, Broxval IM and Saffiotti A, "A navigation
speed using LIDAR.", In: Proceedings of the E3S
system for automated loaders in underground mines",
Web of Conferences, Prague, Czech Republic: E3S
Proceedings of the 5th International Conference on
Web of Conferences; 2019. pp. 1- IO, 2019.
Field and Service Robotics. Port Douglas, Australia:
Springer; 2006. pp. 129-40, 2006. 21. Akai N, Morales LY, Yamaguchi T, Takeuchi E,
Yoshihara Y and Okuda K. "Autonomous driving
14. Fu H, Ye L, Yu Rand Wu T, " An efficient scan-to-
based on accurate localization using multilayer
map matching approach for autonomous driving'',
LiDAR and dead reckoning", IEEE 20th Intelligent
lEEE International Conference on Mcchatronics
Transportation System (ITSC). Yokohama, Japan:
and Automation. Harbin, China: IEEE; 2016. pp.
IEEE; 2018, pp. 1--6, 2018.
1649-54.
22. S. Thrun, W. Burgard and D. Fox., " Probabilistic
IS. T. Berglund, A. Brodnik, H. Jonsson, M. Staffanson
robotics", MlTPrcss, Cambridge, 2005.
and l. Soderkvist, " Planning smooth and obstacle
avoiding B-spline paths for autonomous mining 23. Z. Ren, L. Wang and L. Bi., "Robust GICP-based
vehicles", IEEE Trans Autom Sci Eng, vol. 7 (1), 3D LiDAR SLAM for underground mining
pp. 167-172, 2010. environment. Sensors", vol. 19(1 3), pp. 2915, 2019.
16. H. Kim and Y. Choi, " Development of a LiDAR 24. D. Ghosh, B. Samanta and D. Cbakravarty, "Multi
sensor-based small autonomous driving robot for sensor data fusion for 6D pose estimation and 3D
underground mines and indoor driving experiments", underground mine mapping using autonomous
J Korean Soc Miner Energy Resour Eng, vol. 56, pp. mobile robot", Int J Image Data Fusion, vol. 8 (2),
407-415, 2019. pp. 173-187, 2017.
17. H. Kim and Y. Choi, "Field experiment of a LiDAR 25. BakerC, Morris A, Ferguson D, Thayer S, Whittaker
sensor-based small autonomous driving robot in an C and Omohundro Z., "A campaign in autonomous
underground mine", Tunn Undergr Sp, vol. 30, pp. mine mapping. In: IEEE lnternational Conference
76-86, 2020. on Robotics and Automation. New Orleans, New
York. USA: IEEE; 2004, 2004.
18. S.K. Singh, S. Raval and B. Banerjee, "A robust
approach to identify roof bolts in 3D point cloud 26. N. Evanek, B. Slaker, A. laonacchione and T. Miller,
data captured from a mobile laser scanner", Int J "LiDAR mapping of ground damage in a heading
Min Sci Tech, vol. 31(2), pp. 303-312. 2021. re-orientation case study", Int J Min Sci Tech, vol.
31 ( I), pp. 67-74, 2021.
19. Heonmoo Kim and Yosoon Choi, "Location
estimation of autonomous driving robot and 3D

II The Institution of Engineers (India)