Available online at www.sciencedirect.

com

ScienceDirect
Procedia CIRP 118 (2023) 110–115

16th CIRP Conference on Intelligent Computation in Manufacturing Engineering, CIRP ICME ‘22, Italy

Smart Maintenance Architecture for Automated Guided Vehicles

Dionis Bozhdaraja, Dominik Luckea,b,*, Johannes L. Joostec
a
Hochschule Reutlingen, ESB Business School, Alteburgstraße 150, 72762 Reutlingen, Germany
b
Fraunhofer Institute for Manufacturing Engineering and Automation IPA, Nobelstraße 12, 70569 Stuttgart, Germany
c
Department of Industrial Engineering, Stellenbosch University, Joubert Street, Stellenbosch 7600, South Africa

* Corresponding author. Tel.: +49-7121-271-5005; fax: +49-7121-271-90-5005. E-mail address: [email protected]

Abstract

The increasing complexity and need for availability of automated guided vehicles (AGVs) pose challenges to companies, leading to a focus on
new maintenance strategies. In this paper, a smart maintenance architecture based on a digital twin is presented to optimize the technical and
economic effectiveness of AGV maintenance activities. To realize this, a literature review was conducted to identify the necessary requirements
for Smart Maintenance and Digital Twins. The identified requirements were combined into modules and then integrated into an architecture. The
architecture was evaluated on a real AGV on the battery as one of the critical components.
© 2023 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientific committee of the 16th CIRP Conference on Intelligent Computation in Manufacturing Engineering

Keywords: Smart Maintenance, Automated Guided Vehicle, Digital Twin

1. Introduction addition, unpredictable failures of AGVs cause high costs due

to production losses [5]. The implementation and use of a
In times of digitalization and the shift towards Industry 4.0, digital twin have a high potential to enable greater efficiency,
automated guided vehicles (AGV) are a fundamental quality and transparency as well as a reduction of risks in
component in a smart factory to ensure an automated and intralogistics and therefore in the use of AGVs through
flexible flow of goods in intralogistics [1,2]. The challenges visualizations, analyses and forecasts as well as optimization
posed by increasingly variable products in small batch sizes and [9,10]. Isolated solutions are often used for maintenance [7].
constantly changing production conditions, which demand just- Currently, there is no smart maintenance architecture that
in-time provision of goods, can be solved with the help of ensures an overall view of maintenance for AGVs. This paper
AGVs [3]. The functionality of AGVs plays an important role presents a smart maintenance architecture that ensures an
in intralogistics in the course of Industry 4.0 and the overall view of maintenance for AGVs and allows self-steering
development of a smart factory [4]. However, the efficiency of maintenance functions.
AGVs cannot yet be fully exploited today [5]. One reason is
that today's systems are not able to react automatically to 2. Related Work
failures [5]. Downtimes of AGVs have high failure costs when
causing downtimes in production systems [5]. The maintenance Various definitions can be identified in relation to smart
of AGVs faces major challenges. AGVs consist of many maintenance. According to Henke et al. „Smart Maintenance
components, have a complex structure and therefore require a refers to a learning-oriented, self-regulated maintenance with
considerable amount of time to repair failure and malfunctions the aim of maximizing the technical and economic effectiveness
[5,6]. One of the reasons for this is that maintenance measures of maintenance measures, taking into account the respective
for AGVs are often reactive, as already described, and suitable existing production system, by using digital applications” [7].
maintenance personnel are not directly available [7,8]. In

2212-8271 © 2023 The Authors. Published by Elsevier B.V.

This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientific committee of the 16th CIRP Conference on Intelligent Computation in Manufacturing Engineering
10.1016/j.procir.2023.06.020
 Dionis Bozhdaraj et al. / Procedia CIRP 118 (2023) 110–115 111

According to Bokrantz et al. smart maintenance consists of four controlling the AGV by the digital twin, the provision of
dimensions [11]. These are data-driven decision-making, information from the digital twin to the human, and the
human capital resource, internal integration and external possibility of interaction between the human and the digital
integration [11]. The dimensions are interlinked and support twin. It consists of the functional modules, which are presented
each other. Henke et al. identify six fields of action in relation in the following sections.
to smart maintenance [7]. These include the following: Joint
planning of all actors, availability-oriented maintenance, 3.1. Module AGV
flexible acting and reacting to changes, knowledge
management, requirements-based spare parts management and The AGV module is a partial component of the physical
value contribution of maintenance [7]. world and fulfills function for physical data acquisition as well
To build a smart maintenance architecture, it is suitable to as for control. The data acquired are highly depend on the
use the concept of the digital twin, as this enables a connection structure and components of the AGV. Therefore, critical
between the virtual and physical worlds and can control each components along the components of the AGV type are
other [7,12]. There are different approaches for the identified and evaluated. Typically, these are the wheels, drive
implementation of Architecture. Lee et al. define a five-stage motor, steering drive, brakes, battery, etc. Suitable
architecture with a sequential workflow [13]. According to measurement parameters are determined that, considered
Ashtari Talkhestani et al., the first three stages smart individually or in combination with each other, provide
connection, data-to-information conversion and cyber level of information about failures and maintenance requirements.
this architecture are necessary to realize a digital twin [14]. Once the measurement parameters have been identified, it must
Redelinghuys et al. form a six-stage architecture consisting of be determined whether they can be collected directly from the
sensors, data sources, local data repositories, IoT gateway, measurement equipment of the AGV or whether additional
cloud-based information repositories, simulations, and suitable measurement equipment must be installed in order to
emulations [15]. Here, humans are considered in the obtain real data from the AGV. These data can be e.g.
architecture. The human can receive information and interact odometry data, forces, electrical voltages and temperatures.
with individual elements of the architecture [15]. Ashtari Also, it comprises an interface such as the robot operating
Talkhestani et al. additionally describe the need for system (ROS) standard through which the AGV can be
synchronization and co-simulation interfaces in their controlled for maintenance purposes externally, in order
architecture [14]. They also mention the necessity of using implement a self-steering maintenance function.
models to enable a holistic representation and behavior of the
real object [14]. In their architecture, each digital twin is 3.2. Module existing IT system
assigned a unique ID [14]. Bevilacqua et al. point out the need
for continuous communication between the physical object and The existing IT system module (EITS) represents the
the virtual object [16]. They also see the need to implement function to capture data and information from existing IT
functions such as anomaly detection and decision support in the systems such as file server, databases or the computerized
architecture [16]. Qi et al. describe five essential components maintenance management system. These can be, for example,
that are necessary for the construction of a digital twin [17]. process, geometry, order, material and inventory data.
These are the ability to perceive the physical world, to capture, Furthermore, models that describe the structure and behavior
process, and transmit data, connections within the physical of the AGV are to be transferred to the digital model module.
world, connections between the physical and virtual worlds, These include electrical schematic, geometry, and simulation
and the ability to implement models and services [17]. and emulation models. Data and information such as process,
material and inventory data contained in machine-readable
3. Smart maintenance architecture for AGVs form are transferred to the data acquisition and processing
module for processing. The EITS module is assigned to the
Core idea is to develop a smart maintenance architecture that virtual world. It has a direct link to the data acquisition and
ensures an overall view of maintenance for AGVs and enable processing module and the digital model.
self-steering maintenance functions. The basic architecture
represents the basic framework for the subsequent structure of 3.3. Module data acquisition and processing
the functions within the basic architecture. The basic
architecture is intended to show the layers of the architecture. The task of the data acquisition and processing module
It could be determined that two worlds must be considered for (DAPM) is to acquire, transform, filter, standardize, cleanse
the realization of the architecture. These are the physical world and merge data from the physical world as well as from existing
and the virtual world (Fig.1). The physical world contains the IT systems. The data must be machine-readable and
AGV and the maintenance worker. The virtual world consists unambiguously assignable. In practice, it is therefore necessary
of the digital twin and the communication, connecting both to capture ID and time stamps. This allows measurement data
worlds. Communication is continuous and takes place both from different measurement devices with different
within the digital twin and between the physical world and the measurement intervals and data formats to be transformed,
virtual world. Besides that, four main communication levels are merged and standardized. The transformation should enable the
relevant for the communication between the two worlds. These data to be available in a uniform format. This is a basic
are data acquisition from the physical world, the possibility of requirement for the implementation of the module. The
112 Dionis Bozhdaraj et al. / Procedia CIRP 118 (2023) 110–115

Physical World Virtual World

Msync Async SyncM

AGV EITS Digital Twin

DAPM Database Digital Model
DT1 ITS1
DT1 RM1 … RMn
…
… APM1 …
ITSn … DTn SBM1 … SBMn

DTn APMn

A1 An AnalysisM
ECM
CM1
IM Smart Maintenance Planning
Human IPM …
IM1 KMAG FM RP MP SPM
CMn
PM1 … KM1 FM1 RP1 MP1 SPM1

… … … … … …
IMn
KMn FMn RPn MPn SPMn
PMn

Legend:
AGV: Automated Guided Vehicle DTn: Data Type
EITS: Existing IT System RMn: Representation Model
DAPM: Data Acquisition and Processing Module SBMn: Simulation and Behavioural Models
SyncM: Synchronizaton Module AnalysisM: Analysis Module
Msync: Manual Synchronization KMAG: Knowledge Management and Action Guidelines
Async: Automatic Synchronization FM: Failure Management
IM: Information Module RP: Resource Planning
ECM: External Control Module MP: Maintenance Planning
IPM: Information Provision Module SPM: Spare Parts Management
An: Analysis CMn: Control Module
PMn: Presentation Module ITSn: IT System
APM: Acquisition and Processing Module
Information Transmission Information Transmission with Interaction

Fig. 1. Functional Model of Smart Maintenance Architecture for AGV.

filtering of the data enables extraction of relevant data. The

transformation, filtering and standardization initiate the step of 3.5. Module digital model
data cleansing, in which the identification of missing
measurement data or measurement errors can be specifically The digital model has the task of showing a holistic view of
cleansed. In addition, it is to be checked whether the same IDs the AGV. On the one hand, it enables simulations that show the
occur more than once. Another task of the module is to provide behavior of the AGV under different conditions. In addition,
the processed data to the interfaces to the database module. The other models such as geometry and electrical circuit diagram
data acquisition and processing module is assigned to the models are stored. To enable a holistic description of the AGV,
virtual world. It has a direct connection to the AGV, EITS and representative models are also required. These provide
database modules. information, for example, about the current purpose of the
AGV in the organization, potential further uses, expansion
3.4. Module database possibilities of the AGVs through, e.g. a trailer. The module is
assigned to the digital twin in the virtual world. The module
The database module stores the extracted and processed has a connection to the EITS, database, analysis module and
data. Thereby, both historical and real-time data are stored. the information provision module.
This makes it possible to compare states of the AGV from the
past with states from the present in order to create the basis for 3.6. Module synchronization
predictions about future maintenance measures in the smart
maintenance planning module. This module also has the task The synchronization module is intended to enable
of providing the necessary data to the digital model and synchronization of data and information between the module
analysis module interfaces. The database is connected to the levels. The data and information from the module levels should
data acquisition and processing module, digital model, analysis be synchronized. The time cycle for the exchange is to be
module and the smart maintenance planning module. individually adaptable, so that the data and information
exchange is adjusted and adapted as required.This module
cannot be assigned to a specific entity but is generic and
 Dionis Bozhdaraj et al. / Procedia CIRP 118 (2023) 110–115 113

handles synchronization across modules. To enable the module functions resource planning and spare parts
synchronization, existing interfaces must be used. If no existing management. The information that is necessary from the
interfaces are available, interfaces must be created. It is module function resource planning is the availability of
necessary to define beforehand which data and information will personnel. The spare parts management function checks
be synchronized with which module levels. The whether the necessary spare parts are available. In case that the
synchronization module is a part of the virtual world in the required spare parts are available, the information about the
architecture. The synchronization module is connected with all warehouse location is also transferred. If the spare parts are not
modules. The prerequisite for this module is the creation of a available, information is transferred that shows the necessary
connection for data exchange. As long as there is no reliable spare parts as well as contact data of the supplier, the costs for
connection, no synchronization can take place. It must be the spare parts and the estimated delivery time. This
possible to repeat the connection establishment at regular information from the individual module functions is combined
intervals. This module offers the possibility of manual and transferred to the information module to generate an order.
synchronization or automated synchronization. In addition, checklists, maintenance intervals/dates, inspection
plans, components overview, manufacturer documentation,
3.7. Module smart maintenance planning internal documentation and previous maintenance reports can
be accessed interactively. These should be customizable so that
The smart maintenance planning module has the task of checklists, maintenance reports, internal documentation and
identifying, with the help of various module functions, whether maintenance intervals as well as data from maintenance staff
maintenance measures of an AGV can be remedied by the can be edited and modified by the information provision
vehicle itself or whether the intervention of a maintenance module. In addition, there is a direct link from the database
worker is necessary. Depending on the two different situations, module to the smart maintenance planning module in order to
suitable information for the maintenance activities is to be transfer the latest data, such as availability of personnel in the
provided so that the information module can generate a resource planning function or availability of spare parts in the
maintenance order either for the AGV or for the maintenance spare parts management module function.
personnel. The smart maintenance planning module comprises The resource planning module function has the task of
interconnected functions for maintenance planning, resource managing information on maintenance staff. It records whether
planning, knowledge management and action guidelines, the maintenance personnel are available, what role he plays in
failure management and spare parts management. the organization and whether he has the necessary
The maintenance planning function is used to determine qualifications to perform the demanded maintenance task.
whether a maintenance activity can be carried out by the The failure management sub module consists of the module
vehicle itself or whether the intervention of a maintenance functions failure classification and failure database. The failure
personnel is necessary. If a maintenance personnel is classification function has the task of classifying failures, e.g.
necessary, a suitable maintenance worker is identified and whether a maintenance task must be remedied by a
informed about the maintenance task. In addition, the maintenance personnel or whether the AGV can carry out the
maintenance worker receive a solution catalog, which is corresponding task on its own. The failure database function
intended to support him in carrying out the maintenance task. stores the failures that have occurred in a database. The failures
To make this possible, the smart maintenance planning module are classified according to type and severity, and assigned a
interacts with the information module. The information module unique ID and timestamp. In addition, the failure database
sends failure information that has currently occurred or will should be expandable by the maintenance personnel with
occur in the future to the maintenance planning module additional information describing the failure and the context.
function. The maintenance planning module compares the The spare parts management sub module consists of the
failure with the failure management module function. Here, the module functions digitized spare parts information and spare
failure is firstly classified. It is checked whether it is a failure parts inventories. The digitized spare parts information
that can be executed directly by the AGV itself via the external function provides the maintenance personnel with information
control module (ECM) or whether a maintenance worker is on spare parts. These are divided into assemblies, individual
required. In the case that the AGV can repair the failure itself, parts or raw materials and provide the necessary information
the module function sends a feedback to the information about costs, suppliers including contact details and the
module that contains information for the order generation. dimensions of the spare parts. The spare parts inventories
Possible failures where self-steering maintenance can be provide information on whether and which spare parts are
applied are wheels, drive motor, steering drive, brakes, battery, available, as well as information on delivery times for
etc. If the failure cannot be repaired by the AGV, information unavailable spare parts. It also shows at which locations the
is compiled via the module functions spare parts management, available spare parts are located. Optionally, this module can
failure management, knowledge management and action be extended so that a purchase requisition is issued
guidelines and resource planning to enable order generation. automatically if required. The smart maintenance planning
This information includes the failure type, time of failure and module is a part of the digital twin digital twin in the virtual
failure severity from the module function failure management world. It is connected to the database, the information module
and solutions of the failure type with action guidelines from the and the information provision module.
module function knowledge management and action
guidelines. In addition, further information is provided from
114 Dionis Bozhdaraj et al. / Procedia CIRP 118 (2023) 110–115

3.8. Module information in offline mode, so that location-independent access is possible.

The information is prepared in such a way that the maintenance
The information module has the task of collecting the data personnel can act intuitively. Necessary activities as well as
from the analysis module and comparing it with the failure correlations are to be identified quickly in order to carry out the
database. Through interaction with the smart maintenance necessary maintenance measures. This is made possible with a
planning module, an order is generated according to the semantic search input function. In addition, it comprises the
situation. This can either be sent to the ECM if the maintenance access to the results of the simulations from the digital model,
can be performed by the AGV or to the information provision the analyses from the analysis module and documents on
module if the maintenance personnel has to perform the previous activities. The third function is the interaction of
maintenance task. The necessary information is obtained by the humans with the module and include documentation functions
smart maintenance planning module so that the order can be of the relevant maintenance activities, such as checklists,
generated and sent. The information necessary for order directly on a mobile device. Furthermore, it is required to
generation is described in the smart maintenance planning enable access to the maintenance planning module in order to
module. It has connections to the analysis module, the smart be able to transfer new failures that have occurred to the failure
maintenance planning and the external control module. The database. The adjustments of the modules digital model and
information module is a part of the digital twin in the virtual analysis are to be made available to a certain assigned role, with
world. which an authorization is necessary around the desired function
to accomplish. Also, the module includes communication
3.9. Module analysis components for internet access, telephony and access to
cameras. The internet access is necessary to receive current
The analysis module has the task of performing information about maintenance orders and to obtain necessary
comprehensive analyses using the collected data from the information that is not directly available via the web browser
database module and the simulated data from the digital model or the intranet, as well as to communicate via e-mail or
module. In this process, the data is first merged. This module messenger. The function of telephoning is to enable internal as
already contains predefined analysis methods. The appropriate well as external communication. With cameras, photos or
analysis method is selected depending on the case to be videos can be made, which will be stored in the documentation.
analyzed. With the help of these analyses, existing failures are Optionally, a function could be added that assigned
to be identified and forecasts of future failures and maintenance maintenance personnel can reject the orders. In addition,
requirements are to be made possible. The results will be maintenance personnel should be able to transfer failures
transferred to the information provision module. Furthermore, directly to the failure database. The information provision
the possibility of a routine performance review by suitable module can be assigned to the virtual world. It has connections
maintenance personnel should be given, where adjustments are to the human, digital model, analysis module, information
made regularly in order to optimize the analyses. On this basis, module and smart maintenance planning.
the information is to be transferred to the information module.
A prerequisite for the success of this module is the selection 4. Prototypical Implementation and Validation
and implementation of suitable analysis methods that are
regularly adjusted with regard to their performance. The The different modules forming the smart maintenance
analysis module has links to the database, digital model, architecture for AGVs have been implemented and validated in
information module and information provision module. It is a a manufacturing lab environment. Here, the Neobotix
part of the digital twin in the virtual world. MMO700 has been selected as AGV platform. In an
experimental setup the AGV was moved under various
3.10. Module external control conditions and the individual implemented modules were
tested for their functionality and performance.
The external control module (ECM) has the task of The AGV module was implemented by identifying critical
controlling the AGV and causes an action in the AGV. These components of the AGV. In addition, relevant parameters were
actions are rule-based and can be implemented for different identified providing information about a potential failure. In the
scenarios. The external control module has connections to the use case the battery was identified as a critical component.
AGV and the information module. It can be assigned to the Accordingly, the prototype implementation of the smart
virtual world. The external control module communicates with maintenance architecture was focused on the battery. The
the AGV and provides it with the necessary information, relevant parameters that were identified are state of charge
enabling self-steering maintenance actions. (SoC), remaining useful lifetime (RUL), energy consumption
(EC) and driving time under given battery level. After the
3.11. Module information provision and interaction identification of the parameters, it was checked whether the
parameters can be acquired directly on the AGV or whether
The information provision and interaction module has three additional measuring equipment is required. Parameters such
main functions. One function is to receive maintenance orders. as uptime, temperature, battery charge level, odometry and
Another function is the provision of information for the voltage were recorded via the internal sensors of the AGV. A
execution of maintenance relevant activities from the current clamp was attached to measure the current and power.
maintenance planning module, which should also be viewable
 Dionis Bozhdaraj et al. / Procedia CIRP 118 (2023) 110–115 115

The data acquisition and processing module was industry and how well the smart maintenance architecture
implemented in an automated and rule-based manner, so that performs in the industry.
data sets that are available in different data formats are
converted. In addition, different measurement intervals are
normalized, cleaned and merged into a uniform format. This References
module was realized using Pycharm. In the digital model, a
[1] Andreasson H, Bouguerra A, Cirillo M, Dimitrov DN, Driankov D,
simulation was created using Matlab to simulate the behavior
Karlsson L, Lilienthal AJ, Pecora F, Saarinen JP, Sherikov A, Stoyanov
of the battery under different conditions. In addition, the key T. Autonomous Transport Vehicles: Where We Are and What Is
figures already described were calculated. Missing. IEEE Robot. Automat. Mag; 2015;22(1):64–75.
A Neo4j NoSQL graph database was used to implement the [2] Cupek R, Drewniak M, Fojcik M, Kyrkjebø E, Lin JC-W, Mrozek D,
data storage. Within the graph database, various nodes and Øvsthus K, Ziebinski A. Autonomous Guided Vehicles for Smart
Industries – The State-of-the-Art and Research Challenges. In:
labels were created and connected. These were created for
Krzhizhanovskaya VV, Závodszky G, Lees MH, Dongarra JJ, Sloot
knowledge management and action guidelines, fault PMA, Brissos S et al., editors. Computational Science – ICCS 2020.
management, resource planning, and spare parts management. Cham: Springer International Publishing; 2020. p. 330–343.
Individual links were stored in the graph database that can be [3] European Commission: A Manufacturing Industry Vision 2025.
accessed. This guides maintenance personnel to the documents. European Commission (Joint Research Centre) Foresight Study.
Brussels; 2013.
In addition, an interface was created that allows data from the
[4] Ullrich G, Albrecht T. Fahrerlose Transportsysteme. Wiesbaden:
simulation to be transferred to the graph database. The graph Springer Fachmedien Wiesbaden; 2019.
database serves both as a data repository and as a smart [5] Küster B. Fallbasiertes Expertensystem zur automatisierten Reaktion auf
maintenance planning module, as the individual links within Betriebsstörungen in frei navigierenden fahrerlosen Transportsystemen.
the nodes contain instructions that show the maintenance Hannover; 2020.
[6] Matyas K. Instandhaltungslogistik: Qualität und Produktivität steigern.
personnel exactly which documents or steps need to be
7th ed. München: Hanser; 2019.
completed and executed. In addition, further information about [7] Henke M, Heller T, Stich V, Förster F, Rademacher R, Wolny M, Birtel
the storage of spare parts and tools as well as supplier F, Defér F (Eds.). Smart Maintenance: Der Weg vom Status quo zur
information is included. Furthermore, it is possible to directly Zielvision Projektlaufzeit: 08/2018-09/2019: München: utzverlag; 2019.
view which maintenance employees have already assisted in [8] Yan R, Dunnett SJ, Jackson LM. Novel methodology for optimising the
design, operation and maintenance of a multi-AGV system. Reliability
the rectification of specific failures. Within the database,
Engineering & System Safety; 2018;178:130–9.
interlinks have been implemented that allow easy completion [9] Klein M, Maschler B, Zeller A, Ashtari Talkhestani B, Jazdi N, Rosen R,
of the necessary maintenance documents via Google Docs as Weyrich M. Architektur und Technologiekomponenten eines digitalen
file storage. The information module was implemented by a Zwillings. https://www.researchgate.net/publication/334251026_-
database query, which time-dependently queries individual Architektur_und_Technologiekomponenten_eines_digitalen_Zwillings.
Accessed on 30.04.2022.
values in the graph database and sends orders to the information
[10] Martínez-Gutiérrez A, Díez-González J, Ferrero-Guillén R, Verde P,
provision module or external control module. The external Álvarez R, Perez H. Digital Twin for Automatic Transportation in
control module has the task of externally controlling the AGV. Industry 4.0. Sensors; 2021;21(10):3344.
For this purpose, a predefined driving job is transmitted to the [11] Bokrantz J, Skoogh A, Berlin C, Wuest T, Stahre J. Smart Maintenance:
AGV via a ROS API. In case of a required maintenance, the an empirically grounded conceptualization. International Journal of
Production Economics; 2020;223:1-17.
external control module generates an order for the driving job,
[12] Neto AA, Deschamps F, Da Silva ER, Lima EP de. Digital twins in
which navigates the AGV to a predefined maintenance area. In manufacturing: an assessment of drivers, enablers and barriers to
this maintenance area, the AGV shuts down, for example, to implementation. Procedia CIRP; 2020;93:210–5.
avoid consequential damage to other assemblies. In this case, [13] Lee J, Bagheri B, Kao H-A. A Cyber-Physical Systems architecture for
the external control module generates an order for the driving Industry 4.0-based manufacturing systems. Manufacturing Letters;
2015;3:18–23.
job, which navigates the driverless transport system to a
[14] Ashtari Talkhestani B, Jung T, Lindemann B, Sahlab N, Jazdi N,
predefined maintenance area. Overall, through the Schloegl W, Weyrich M. An architecture of an Intelligent Digital Twin
implementation, integration and test of the modules the in a Cyber-Physical Production System. at - Automatisierungstechnik;
feasibility of the smart maintenance architecture for AGVs was 2019;67(9):762–82.
confirmed. [15] Redelinghuys A, Basson A, Kruger K. A Six-Layer Digital Twin
Architecture for a Manufacturing Cell. In: Borangiu T, Trentesaux D,
Thomas A, Cavalieri S, editors. Service Orientation in Holonic and
5. Conclusion and outlook Multi-Agent Manufacturing. Cham: Springer International Publishing;
2019. p. 412–423.
This paper presents a smart maintenance architecture for [16] Bevilacqua M, Bottani E, Ciarapica FE, Costantino F, Di Donato L,
AGVs using a digital twin. The human is also part of the Ferraro A, Mazzuto G, Monteriù A, Nardini G, Ortenzi M, Paroncini M,
Pirozzi M, Prist M, Quatrini E, Tronci M, Vignali G. Digital Twin
architecture and interact with the individual modules within the
Reference Model Development to Prevent Operators’ Risk in Process
physical and virtual worlds. The architecture was implemented Plants. Sustainability; 2020;12(3):1088.
and the feasibility of the architecture was confirmed with in a [17] Qi Q, Tao F, Hu T, Anwer N, Liu A, Wei Y, Wang L, Nee A. Enabling
manufacturing lab environment. Future research activities are technologies and tools for digital twin. Journal of Manufacturing
required to determine how well the smart maintenance Systems; 2021;58:3–21.
architecture can be implemented to different AGV types in