THERE’S a variety of data coming from the manu-

facturing site today. Along with the development

of the way various data are processed, we have also been

concerned about the way various data are displayed.

The Manufacturing Execution System (MES) has cre-

ated multiple screens that combine data to show data

from diﬀerent perspectives. This phenomenon has not

necessarily changed to an advanced form. It can be use-

ful for users who want to see and analyze a variety of

data combinations, but there are many inconveniences

for simple information. You need to go to the PC where

the system can be connected and access it. In addition,

even though you can see it on your mobile device, you

need to ﬁnd a menu of many categories and access screens

in subcategories to see the data you want. And to see

simple data, various data are viewed together on the

connected screen, so the viewing speed is delayed, and

the view can be dispersed because it shows data that I

don’t need. To compensate for these points, ChatMES,

a simple manufacturing bot using Chatbot, was studied.

Chatbot is an application that mimics conversations

with people and allows them to interact with virtual

applications that seem like they’re actually talking [1].

Chatbots can be as easy as basic applications to answer

simple queries, and they can also require cutting-edge

technology, such as personal assistants, who can use big

data to answer queries in a variety of casesBut in manu-

facturing, AI wasn’t very involved in chatbots [1]. Today,

most services have automated chatbots that can interact

with users [3]. Automation through chatbots and many

situations demonstrated the usefulness of improving the

user experience [4] The study highlights that chatbot ap-

plications will be a key technology in IT technology in the

future [5]. Robot Process Automation (RPA) technol-

ogy is the technology that makes chatbots like personal

assistants[5]. RPA is the evolution of next-generation au-

tomation systems that deliver faster, more accurate per-

formance and higher return on investment in business

applications [6].

In this study, we designed a chatbot that works with

MES and uses the technology of RPA. With the devel-

opment of smart manufacturing, processes and inquiry

screens through screens are becoming more complicated

in a manufacturing environment with a lot of data. In

Chatmes, you can get the information you want any-

time, anywhere through an accessible messenger, and

you can work through messenger by registering simple

queries, scheduling tasks, and even complex tasks with

intent through RPA. In addition, Chatmes designed an

architecture that enables RPA by accessing various mes-

senger programs and legacy systems except MES with

the API plug-in layer in mind.

This study consists of: Section II. describes the rel-

evant operation. Section III. presents the architecture of

the manufacturing chatbot and describes the architec-

ture. Section IV. conﬁrms the results of the manufac-

turing chatbot through experiments. Finally, section V.

discusses conclusions and future research plans.

Smart manufacturing is a term coined by several or-

ganizations, such as the Department of Energy (DoE)

and the National Institute of Standards and Technol-

ogy (NIST) in the United States. Using information and

communication technology (ICT) and data analysis, the

manufacturing site was improved and the operation of

the factory was emphasized [7, 8, 9] Smart manufactur-

Design and Implementation of RPA Based ChatMES System

Architecture for Smart Manufacturing

SANGHOON DO, JONGPIL JEONG

Department of Smart Factory Convergence, Sungkyunkwan University 2066 Seobu-ro, Jangan-gu,

Suwon 16419, REPUBLIC OF KOREA

Abstract- Today, with the generation of a lot of data in the manufacturing field, research is being

actively conducted on how data is processed. But we’re building more complex screens to process and

show a lot of data. However, there are many times when manufacturing sites want fast, concise data.

It is possible through Chatbot with Robot Process Automation (RPA) for checking and processing

concise data regardless of time and place. Provides an architecture that can handle complex system

screen queries and complex processes.

Keywords- RPA, Chatbot, Smart Manufacturing, MES

Received: April 7, 2021. Revised: June 15, 2022. Accepted: July 11, 2022. Published: September 13, 2022.

1. Introduction

2. Related Work

2.1 Smart Manufacturing

WSEAS TRANSACTIONS on COMPUTER RESEARCH

DOI: 10.37394/232018.2022.10.11

Sanghoon Do, Jongpil Jeong

E-ISSN: 2415-1521

Volume 10, 2022

ing integrates various technologies such as CyberPhysical

Production Systems (CPPS), IoT, robotics/automation,

big data analytics, and cloud computing to enable data-

driven connectivity [10]. Intelligent manufacturing has

been used in the same sense as smart manufacturing. The

technology and essential elements of smart manufactur-

ing were studied to explain the diﬀerence between smart

manufacturing and intelligent manufacturing [7, 11, 12].

Research has been conducted on the features and es-

sentials of smart manufacturing and intelligent manufac-

turing technology [13]. Intelligent manufacturing focuses

more on the technical side and less on the organizational

side than on smart manufacturing. In the context of In-

dustry 4.0, intelligent manufacturing is drawing tremen-

dous attention from government, corporate and academic

researchers [14].

Chatbots are also called messenger bots because they

are based on text messages exchanged between users and

bots in chat spaces. When a user enters a text message

in a virtual space, the software analyzes the message,

creates an appropriate response, and sends the message

back to the user. Therefore, the user can communicate

with the soft-agent-bot through interaction with the soft-

agent-bot and verify that the soft-agent-bot has per-

formed the correct operation. In general, chatbots are

systems that send corresponding messages according to

already deﬁned rules as needed. In the initial develop-

ment version, a method of checking simple matches of

keywords and return values was performed. However,

due to recent advances in various technologies, the re-

quirements are deﬁned by identifying and analyzing the

natural language entered by users. Chatbots can fulﬁll

requirements and respond to results [15]. The function

of these chatbots means that they can act as agents,

as mentioned above. The agent is an abstract object

in terms of software. Engineering has the same area as

methods, functions, and modules, but the diﬀerence is

that it performs autonomous operations instead of sim-

ple functions [16, 17].

Process automation technology is used to automati-

cally control processes. Technological advances have en-

abled industrial automation systems to be introduced

wherever possible. Process automation can be divided

into hard automation and soft automation [18]. Hard au-

tomation refers to the performance of ﬁxed iterations on

a given product, and soft automation performs a vari-

ety of non-ﬁxed operations. RPA is a technology that

belongs to soft automation[19] The software bot handles

data (read, write, numerically, and compute) with ap-

plications through procedured workﬂows [20]. In various

cases of analyzing RPA inside the device, it was con-

ﬁrmed that the processing speed of some tasks in the or-

ganization was shortened and productivity was increased

when RPA was applied. In a recent study, a prototype of

an interactive digital assistant based on natural language

processing has been presented to present tasks related to

intelligent RPA. Through this, it is predicted that inter-

active bot will be able to solve business [21, 22].

In our study, ChatBot’s research using RPA was con-

ducted so that the current status of the manufacturing

site can be checked quickly and easily anywhere.

Fig. 1: ChatMES System Architecture

Figure 1 shows the architecture we studied. All pro-

cessing is done on the Bot Platform. The Bot Platform

is divided into the Management Layer, Messenger I/F

Layer, Adapter Layer, Business Layer, and API Gate-

way Layer.

First, in the Management Layer, there are Man-

age Users, Manage Bots, Manage Feedback, and Man-

age Statistical. Manage Users manages registered users.

Only authorized users are allowed to use it, and be-

cause they have permission to view information based

on their ratings, they are granted permissions based on

their query intentions. Manage Bots manages data that

2.2 Chatbot

2.3 RPA

3. ChatMES System Architecture

Based on RPA

WSEAS TRANSACTIONS on COMPUTER RESEARCH

DOI: 10.37394/232018.2022.10.11

Sanghoon Do, Jongpil Jeong

E-ISSN: 2415-1521

Volume 10, 2022

can be viewed. Data that users use a lot is created and

managed by intention. Manage Feedback veriﬁes that the

user’s intended data has been sent, analyzes messages

that have not been identiﬁed, and ﬁne-tunes them ac-

cording to their intentions by type of message to modify

their intentions in Manage Data. Manage Statistical is to

understand the statistics of the messages you use. The

high frequency of messages is modiﬁed to be more diver-

siﬁed and used, and the low frequency of use is further

supplemented to increase the frequency of users’ use so

that they can be used more conveniently.

Second, Messenger I/F Layer. You will use it to con-

nect your messenger to your Chatbot. It helps you com-

municate in both directions and is used for smooth mes-

sage delivery. In this study, Knox Messenger was used,

but in order to use various messengers, API can be con-

nected to the layer in a plug-in manner.

Third, the Adaptor Layer. This layer is associated

with the Messenger I/F layer to receive a message

through the Incoming Adapter and send the message to

the Business Layer. In contrast, the Outgoing Adapter

receives outgoing messages from the Business Layer and

sends them to the Messenger I/F Layer. Although it ap-

pears to be two-way communication on the layer, it is

divided internally because the communication direction

of the incoming and outgoing messages is diﬀerent.

Fourth, the Business Layer includes Message Analy-

sis, Manage Data, Manage Message, and Response Pro-

cessg. Message Analysis preprocesses messages received

through the Incoming Adapter and distributes the types

and variables by intent. Identify and distinguish inten-

tions from pretreatment. Find the right intention in Man-

age Data with diﬀerentiated intentions and match the

variables accordingly. Manage Message uses JSON to

separate the instant and scheduled messages from the

results data received and to manage them in a message

queue. Response Processing generates a result message

with the data stored in the message queue. And leave

the message history and logs so that you can verify that

it worked properly.

Fifth, API Gateway Layer exists to query data and

receive results from the system. Currently, it is designed

to receive only MES data, so it connects to the MES Sys-

tem using the JSON API for JSON connectivity. How-

ever, if users want to connect to other legacy systems

such as ERP and SCM, not just MES systems, they

can connect to non-MES systems and receive messages

by adding APIs of promised communication to the API

Gateway Layer.

We experimented with an architecture designed in

Figure 1. ChatBot was designed through the Brity RPA

of Samsung SDS, an open RPA system, and Samsung

SDS used Knox Messenger.

igure 2 shows that the message is communicated in

JSON format and that the variable is received and pro-

cessed. Messages will be received through the Incoming

Adaptor in the Adaptor Layer. A message is received

Fig. 2: Message Communication

from the Business Layer, preprocessed, matched with the

intention of the Mange Date, and processed by the ap-

propriate application in the MES system to query the

promised data. The data received by messenger was di-

vided into date, intention, and line. The date was set as

a mandatory variable because there is a risk that a large

amount of data will be inquired. If you do not enter a

date, it will be recognized as a null value, so you will

query the date you sent the message, and if you do not

specify a line, the MES system will query all lines. You

will receive a result message through JSON through the

server. The body part of the result message is received

and the ﬁrst value indicates the value of the intention

inquired. 1 is a message about the production status and

will reply according to the format. If you reply to the sec-

ond value in a table format, the subsequent value sends

the result values that you see in order. Response Pro-

cessing generates a message in a format that responds to

the user via the Outgoing Adaptor and Messenger API.

Fig. 3: Massage of Production Progress

Figure 3 received a message of line status, one of the

intentions registered in Chatbot, through messenger to

check the production progress of line 1. You have not

speciﬁed a date, so you have received a production status

of 1 line on the day you sent the message.

Figure 4 is a simple RPA that is set to communicate

the intent of the 2-line defect status to the messenger.

Figure 5 is the result of experimenting with a message

to check the bad status of the 2 line. I sent a message to

Chatbot according to the intention of the defect status

for each registered line and received the result of the

defect status on the day of inquiry of the 2nd line through

RPA.

4. Experiment and Results

WSEAS TRANSACTIONS on COMPUTER RESEARCH

DOI: 10.37394/232018.2022.10.11

Sanghoon Do, Jongpil Jeong

E-ISSN: 2415-1521

Volume 10, 2022

Fig. 4: Scheduling of Defect

Fig. 5: Massage of Defect

Through the results of the experiment, I was able to

quickly and easily receive the data of the designated in-

tention through Chatbot RPA using messenger. Apply-

ing RPA to messenger has four advantages. First, there

is a low dependence on the type of device. The form

of messenger exists everywhere in the form of a PC or

mobile. Even now, wearable devices can also function as

messengers. The advantage is that authorized users can

receive MES information from anywhere. You can receive

information from anywhere in the factory or outside the

factory. Second, it has the advantage of being able to

access multiple messengers and multiple legacy systems

by adding APIs to Messenger I/F Layer and API Gate-

way Layer in the architecture. Third, there is no need

for equipment redevelopment. In the past, the develop-

ment of the same screen was diﬀerent due to the size of

the device in the MES system, so the development of the

PC version and the mobile version were renewed. How-

ever, there is no need to develop new information through

messenger because the layout of messenger is the same

regardless of device. Fourth, you don’t have to perform

many steps to access the system screen for information

inquiry, and you don’t have to display standardized data

that you don’t need.

In this study, we proposed an architecture of

ChatMES that allows complex messengers to query and

process information from manufacturing sites in a simple

and eﬃcient manner. We presented the architecture to

connect messenger and MES system and considered the

management aspect to eﬃciently manage Chatbot. This

architecture is highly scalable with an architecture that

considers connecting any variety of messengers to any

variety of legacy systems. In addition, the MES charac-

teristics are directly related to the manufacturing site,

so we designed the intent to reduce the load and imple-

mented an architecture that can be performed according

to the designated intent. There have been fewer studies

using chatbots in closed manufacturing systems than in

other ﬁelds. I hope that this study will help to further

activate the use of chatbots using RPA in the manufac-

turing site.

Future research needs to be done on natural language.

There are diﬃculties in management because we made

movements according to the exact intentions. Research

on natural language is needed to enable interaction, and

we will conduct research so that the standardized screen

of MES can be reduced and the UI of MES can be por-

talized.

This research was supported by the SungKyunKwan

University and the BK21 FOUR (Graduate School Inno-

vation) funded by the Ministry of Education (MOE, Ko-

rea) and National Research Foundation of Korea (NRF).

[1] T. -Y. Chen, Y. -C. Chiu, N. Bi and R. T. -

H. Tsai, ”Multi-Modal Chatbot in Intelligent Man-

ufacturing,” in IEEE Access, vol. 9, pp. 82118-

82129, 2021, Available: https://doi.org/10.1109/

ACCESS.2021.3083518

[2] H. Chen, X. Liu, D. Yin and J. Tang, ”A survey on di-

alogue systems: Recent advances and new frontiers”,

ACM SIGKDD Explor. Newslett., vol. 19, no. 2, pp.

25-35, Nov. 2017. Available: https://doi.org/10.

1145/3166054.3166058

[3] B. Nardi, S. Whittaker and E. Bradner, ”Interaction

and Outeraction: Instant Messaging in Action”, Proc.

of the 3rd CSCW Conference, pp. 79-88, 2000. Avail-

able: https://doi.org/10.1145/358916.358975

[4] L. Klopfenstein, S. Delpriori, S. Malatini and A.

Bogliolo, ”The Rise of Bots: a Survey of Conversa-

tional Interfaces Patterns and Paradigms”, Proc. of

the 12th DIS Conference, pp. 555-565, 2017. Avail-

able: https://doi.org/10.1145/3064663.3064672

[5] The Road to Enterprise AI, RAGE Frameworks,

2017.

[6] Lacity, Mary C., and Leslie P. Willcocks. ”A new

approach to automating services.” MIT Sloan Man-

5. Conclusions

Acknowledgment

References

WSEAS TRANSACTIONS on COMPUTER RESEARCH

DOI: 10.37394/232018.2022.10.11

Sanghoon Do, Jongpil Jeong

E-ISSN: 2415-1521

Volume 10, 2022

agement Review 58.1 (2016): 41-49. Available: http:

//eprints.lse.ac.uk/id/eprint/68135

[7] Klaus-Dieter Thoben, Stefan Wiesner, and Thorsten

Wuest, ““Industrie 4.0” and Smart Manufacturing –

A Review of Research Issues and Application Exam-

ples,” Int. J. Automation Technol., Vol.11, No.1, pp.

4-16, 2017 Available: https://doi.org/10.20965/

ijat.2017.p0004

[8] Martin, Bradley, et al. ”A strategic assessment of the

future of us navy ship maintenance: challenges and

opportunities.” (2017). Available: https://apps.

dtic.mil/sti/citations/AD1085610

[9] Osterrieder, P.; Budde, L.; Friedli, T. The smart

factory as a key construct of industry 4.0: A sys-

tematic literature review. Int. J. Prod. Econ. 2020,

221, 107476. Available: https://doi.org/10.1016/

j.ijpe.2019.08.011

[10] Sj¨odin, D.R.; Parida, V.; Leksell, M.; Petrovic, A.

Smart Factory Implementation and Process Innova-

tion: A Preliminary Maturity Model for Leveraging

Digitalization in Manufacturing Moving to smart fac-

tories presents speciﬁc challenges that can be ad-

dressed through a structured approach focused on

people, processes, and technologies. Res. Technol.

Manag. 2018, 61, 22–31. Available: https://doi.

org/10.1080/08956308.2018.1471277

[11] Wang, B.; Tao, F.; Fang, X.; Liu, C.; Liu, Y.; Frei-

heit, T. Smart Manufacturing and Intelligent Manu-

facturing: A Comparative Review. Engineering 2021,

7, 738–757. Available: https://doi.org/10.1016/

j.eng.2020.07.017

[12] Mittal, S.; Khan, M.A.; Romero, D.; Wuest, T.

Smart manufacturing: Characteristics, technologies

and enabling factors. Proc. Inst. Mech. Eng. Part B J.

Eng. Manuf. 2019, 233, 1342–1361. Available: https:

//doi.org/10.1177/0954405417736547

[13] Zhong, R.Y.; Xu, X.; Klotz, E.; Newman, S.T. In-

telligent Manufacturing in the Context of Industry

4.0: A Review. Engineering 2017, 3, 616–630. Avail-

able: https://doi.org/10.1016/J.ENG.2017.05.

015

[14] J. Wan, M. Yi, D. Li, C. Zhang, S. Wang and

K. Zhou, ”Mobile Services for Customization Man-

ufacturing Systems: An Example of Industry 4.0,”

in IEEE Access, vol. 4, pp. Available: 8977-8986,

2016,doi:10.1109/ACCESS.2016.2631152.

[15] Noguera-Arnaldos, J.´

A., Paredes-Valverde, M.A.,

Salas-Z´arate, M.P., Rodr´ıguez-Garc´ıa, M.´

A.,

Valencia-Garc´ıa, R., Ochoa, J.L.: im4Things: an

ontology-based natural language interface for

controlling devices in the Internet of Things. In:

Alor-Hern´andez, G., Valencia-Garc´ıa, R. (eds.) Cur-

rent Trends on Knowledge-Based Systems, vol. 120,

pp. 3–22. Springer, Cham (2017). Available: https:

//doi.org/10.1007/978-3-319-51905-0_1

[16] Nwana, Hyacinth S. ”Software agents: An

overview.” The knowledge engineering re-

view 11.3 (1996): 205-244. Available: https:

//doi.org/10.1017/S026988890000789X

[17] Huang, Jizhou, Ming Zhou, and Dan Yang. ”Ex-

tracting Chatbot Knowledge from Online Discussion

Forums.” IJCAI. Vol. 7. 2007.

[18] ”UIPath vs Automation Anywhere-Blue

Prism”, UiPath Community Forum -

Robotic Process Automation, 2017, [online]

Available: https://forum.uipath.com/t/

uipath-vs-automationanywhere-blue-prism/

5305

[19] M. Lacity, Robotic Process Automation: The Next

Transformation Lever for Shared Services. Curators’

Professor, College of Business University of Missouri-

St. Louis, 2018.

[20] Jalonen, Helge. ”Assessing robotic process automa-

tion potential.” Tampere University of Technology

(2017).

[21] Aguirre, S., Rodriguez, A. (2017). Automation

of a Business Process Using Robotic Process Au-

tomation (RPA): A Case Study. In: Figueroa-

Garc´ıa, J., L´opez-Santana, E., Villa-Ram´ırez, J.,

Ferro-Escobar, R. (eds) Applied Computer Sci-

ences in Engineering. WEA 2017. Communications

in Computer and Information Science, vol 742.

Springer, Cham. Available: https://doi.org/10.

1007/978-3-319-66963-2_7

[22] S. Chen, B. Mulgrew, and P. M. Grant, “A clus-

tering technique for Y. Rizk, V. Isahagian, S. Boag,

Y. Khazaeni, M. Unuvar, V. Muthusamy, et al., ”A

conversational digital assistant for intelligent process

automation”, Proc. Int. Conf. Bus. Process Manage.,

pp. 85-100, 2020. Available: https://doi.org/10.

1007/978-3-030-58779-6_6

Creative Commons Attribution License 4.0

(Attribution 4.0 International, CC BY 4.0)

This article is published under the terms of the Creative

Commons Attribution License 4.0

https://creativecommons.org/licenses/by/4.0/deed.en_US

WSEAS TRANSACTIONS on COMPUTER RESEARCH

DOI: 10.37394/232018.2022.10.11

Sanghoon Do, Jongpil Jeong

E-ISSN: 2415-1521

Volume 10, 2022