Emmi Rajala
AUTONOMOUS MOBILE ROBOTS IN SMALL BATCH SIZE PRODUCTION
Master of Science Thesis
Faculty of Engineering and Natural Sciences
Examiners: Professor Minna Lanz, Dr. Eeva Järvenpää
October 2021
Emmi Rajala: Autonomous mobile robots in small batch size production Master of Science Thesis
Tampere University
Master’s Degree Program in Automation Engineering October 2021
This thesis concentrates on autonomous mobile robots (AMRs) in small batch size production.
The main objective was to determine sales arguments for autonomous mobile robot usage in production environment. This objective was approached by finding answers to three research questions about deployment aspects, suitable tasks, and current market state.
The research started by studying literature about autonomous mobile robots, and other sub- jects that were strongly linked to the thesis subject. In addition, the results of prior surveys were studied. Four research methods were used for finding answers to the research questions. Those methods were literature review, case study, reference case review and open survey. All the used methods were found to be suitable for this purpose but delimitation to the small batch size man- ufacturing was proven to be difficult in most of these methods.
By studying mobile robot related standards and other literature, deployment aspects were found and then tested and fulfilled in case study. These two methods were also used for gathering a list of benefits and disadvantages. To find suitable tasks for AMRs in production environment were searched by doing a reference case review. These tasks were then complemented by open survey results. Understanding of the current market state was also gained from the open survey results.
The reliability of the collected research data was discussed, and each result were analyzed and validated using suitable qualitative or quantitative methods. Data-driven analysis was per- formed to the literature review results to form categorized and summarized lists of the deployment aspects and AMR benefits and disadvantages. The case study data was analyzed using deduc- tive method. Descriptive statistical analysis was performed for the reference case review data to form visual graphs. Same data analysis methods were used also for the open survey data also for graph formation. All the validation was done mainly by comparing the results to literature or to recent surveys.
Even though some results could not be generalized, due to limited data and delimitation diffi- culties, useful results were able to achieve. The most significant result was the deployment as- pects because research concentrating on that field seem to be lacking. The deployment aspects were dived into the sales and implementation phase. Many of the found advantages like flexibility and fast payback time are commonly known, but some of the disadvantages were not so common.
The AMRs are not so plug and play type of products that they are generally marketed to be.
Examples of suitable tasks for AMRs were mostly found from automotive and electrical industry and those tasks were mainly production material handling related. The current market seems to be open minded towards AMRs, but the knowledge of AMRs is mainly lacking. Sales arguments were then formed by combining the analyzed answers for the research questions. At the end, conclusions are made.
Keywords: Autonomous mobile robots, flexible manufacturing, adaptive automation, batch production
The originality of this thesis has been checked using the Turnitin OriginalityCheck service.
Emmi Rajala: Autonomiset mobiilirobotit piensarjatuotannossa Diplomityö
Tampereen yliopisto
Automaatiotekniikan diplomi-insinöörin tutkinto-ohjelma Lokakuu 2021
Tässä diplomityössä kekitytään autonomisiin mobiilirobotteihin (AMR:iin) piensarjatuotan- nossa. Työn päätavoitteena on määrittää myyntiargumentteja mobiilirobottiavusteiselle tuotan- nolle. Tavoitetta lähestytään kolmen tutkimuskysymyksen avulla, jotka liittyvät käyttöönotossa huomioitaviin asioihin, soveltuviin tehtäviin ja tämänhetkisen markkinatilanteen määritykseen.
Tutkimus aloitettiin tutustumalla autonomisiin mobiilirobotteihin liittyvään kirjallisuuteen ja mui- hin tähän tutkimukseen liittyviin aihealueisiin. Lisäksi muiden viime aikoina samasta aiheesta teh- tyjen tutkimuksien tuloksiin perehdyttiin. Neljä tutkimusmenetelmää valittiin siten että niillä pysty- tään vastaamaan asetettuihin tutkimuskysymyksiin. Nämä valitut menetelmät olivat kirjallisuus- tutkimus, tapaustutkimus, referenssi tapaustutkimus ja avoin kyselytutkimus. Kaikkien näiden va- littujen metodien todettiin olevan pääosin soveltuvia tähän tarkoitukseen, mutta osassa näistä oli vaikeaa tehdä rajausta piensarjatuotantoon.
Tutkimalla mobiilirobotiikkaan liittyviä standardeja ja muita kirjallisuuslähteitä, käyttöönotossa huomioitavia asioita löydettiin ja niitä testattiin ja tarkennuksia tehtiin tapaustutkimuksessa. Näitä samoja menetelmiä käytettiin myös selvittämään AMR:ien hyödyt ja huonot puolet tuotantokäy- tössä. Jotta löydettäisiin mobiiliroboteille sopivia tehtäviä tuotantoympäristössä, etsittiin mobiiliro- bottimyyjien internetsivuilta referenssitapauksia. Löydettyjä sovelluskohteita vielä täydennettiin avoimesta kyselytutkimuksesta saaduilla tiedoilla. Markkinan nykytilaa pyrittiin vielä selvittämään kyselytutkimuksen voimin.
Kerätyn datan luotettavuutta tarkasteltiin ja kaikille eri tuloksille valittiin sopivat kvalitatiiviset tai kvantitatiiviset analysointi ja validointi menetelmät. Tietoon pohjautuva analyysi suoritettiin kir- jallisuustutkimuksen tuloksille tavoitteena muodostaa luokiteltuja ja tiivistettyjä kokonaisuuksia mobiilirobottien käyttöönotossa huomioitavista asioista, sekä hyödyistä että huonoista puolista.
Tapaustutkimusta analysointiin deduktiivisella menetelmällä. Kuvaileva tilastollinen analyysi teh- tiin puolestaan referenssi tapaus tutkimukselle, jotta tuloksista voitaisiin muodostaa visuaalisia graafeja. Samaa menetelmää käytettiin myös avoimen kyselytutkimuksen tuloksien analysoin- nissa. Kaikki validointi suoritettiin pääosin vertailmalla saatuja tuloksia kirjallisuuteen tai toisten tekemiin viimeaikaisten kyselyjen tuloksiin.
Vaikka osaa tuloksista ei kyetty yleistämään johtuen rajallisesta tutkimuksessa kerätystä da- tasta ja rajauksen hankaluudesta, tutkimuksessa pystyttiin kuitenkin saavuttamaan käyttökelpoi- sia tuloksia. Merkittävin saavutettu tulos oli käyttöönotossa huomioon otettavat asiat, sillä tähän aiheeseen keskittyvää aiempaa tutkimusta ei tuntunut löytyvän. Kerätyt käyttöönotossa huomioon otettavat asiat jaettiin myyntivaiheessa ja varsinaisessa käyttöönotossa huomioitaviin asioihin.
Suurin osa löydetyistä hyödyistä kuten joustavuus ja nopea takaisinmaksuaika, ovat laajalti jo entuudestaan tiedossa, sillä niitä käytetään apuna markkinoinnissa. Huonoista puolista löytyi kui- tenkin yksi yllättävä seikka, joka oli se, että AMR:t eivät tuntuneet olevan niin helposti käyttöön- otettavissa kuin yleisesti annetaan ymmärtää. Esimerkkejä soveltuvista mobiilirobottien tehtävistä tuotanto käytössä löytyi lähinnä auto- ja sähköteollisuuden aloilta. Löydetyt tehtävät lukeutuivat lähinnä tuotannossa käytettävien tuotteiden ja materiaalien kuljetuksiin. Tämänhetkinen markkina tuntuu olevan avoin AMR:lle, mutta yleinen ymmärrys mobiilirobotiikasta tuntuu olevan vielä melko matalla tasolla. Myyntiargumentit koottiin analysoiduista tutkimuskysymysten vastauksista ja lopuksi esitetään vielä johtopäätökset.
Avainsanat: Autonomiset mobiilirobotit, joustava tuotanto, mukautuva tuotanto, piensarjatuotanto
Tämän julkaisun alkuperäisyys on tarkastettu Turnitin OriginalityCheck –ohjelmalla.
This thesis is done for JTA Connection’s sales team. They were in need for sales argu- ments and some marked research for production cells served by autonomous mobile robots. The thesis process started in the beginning of 2021.
First of all, I would like to thank Jarkko Lepistö for giving me this amazing thesis work.
Secondly, I’m very thankful for all the JTA’s personnel that were helping me in this pro- cess. Thank you Mirva Vuosjoki for finding me this subject. I’m also very appreciative for the examiners Dr. Eeva Järvenpää and Professor Minna Lanz for guiding me through this thesis process.
Lastly, I’m very great full for AGCO Power for giving me study leave for finishing my studies and of course to my family for all the support and understanding.
“Don’t let the sun go down without saying thank you to someone, and without admitting to yourself that absolutely no one gets this far alone.”
- Stephen King
In Tampere, 20.10.2021
Emmi Rajala
1. INTRODUCTION ... 1
2.KEY ELEMENTS OF ADAPTIVE AUTOMATION IN SMALL BATCH SIZE PRODUCTION ... 4
2.1 Adaptive automation paradigms ... 4
2.1.1 Flexible manufacturing systems ... 4
2.1.2 Reconfigurable manufacturing systems ... 6
2.2 Production management ... 7
2.2.1Material management ... 7
2.2.2Material flow ... 8
2.3 Cell layout configuration ... 10
2.4 Autonomous mobile robots ... 12
2.5 Deployment aspects for autonomous mobile robots ... 15
2.5.1 Deployment aspects from standards ... 15
2.5.2Deployment aspects from literature ... 17
2.6 Benefits and disadvantages of AMRs in production environment ... 20
2.7 Technology acceptance ... 22
2.8 Fourth industrial revolution ... 25
3. RESEARCH DESCRIPTION AND IMPLEMENTATION ... 28
3.1 Research methodology ... 28
3.2 Literature review for deployment aspects, benefits, and disadvantages 30 3.2.1 Standards review for deployment aspects ... 30
3.2.2 Publication review for design aspects ... 31
3.2.3 Publication review for benefits and disadvantages ... 33
3.3 Case Study ... 33
3.3.1Starting points for case study ... 34
3.3.2Current state of case study company after AMR deployment ... 35
3.3.3Data collection methods in case study ... 37
3.4 Reference case review for AMR usage in production ... 39
3.5 Open survey ... 41
3.5.1Starting points for customer survey ... 42
3.5.2Survey implementation process ... 42
3.6 Data analysis methods and process description ... 44
3.7 Validity and reliability methodology ... 46
4. RESEARH RESULTS FROM DIFFERENT METHODS ... 48
4.1 Literature review results ... 48
4.1.1 Deployment aspects from literature ... 48
4.1.2 Benefits and disadvantages found from literature ... 49
4.2 Case study results ... 50
4.2.3 AMR capacity measurements ... 53
4.2.4 Return on investment and payback calculations ... 54
4.3 Reference case review results ... 57
4.4 Open survey results ... 60
5. RESEARCH RESULT ANALYSIS AND EVALUATION ... 68
5.1 Analysis of the research results ... 68
5.1.1 Analysis of the literature review results ... 68
5.1.2 Analysis of the case study results ... 71
5.1.3 Analysis of the reference case review ... 73
5.1.4 Analysis of the open survey results ... 74
5.2 Reliability and validity analysis ... 77
5.2.1Literature review validation and reliability discussion ... 77
5.2.2Validation of the data collected from case study and reliability discussion ... 78
5.2.3Reference case review data validation and reliability discussion . 78 5.2.4 Validation of the collected survey data and reliability discussion . 80 5.3 Evaluation of the overall research process ... 83
5.4 Future research proposal ... 84
6. KEY FINDINGS ... 87
6.1 Aspects to consider when planning AMR deployment in production environment ... 87
6.2 Suitable tasks for AMRs in production environment ... 88
6.3 Current market state of AMRs in manufacturing industry ... 89
6.4 Sales arguments for AMR assisted production ... 90
7. CONCLUSIONS ... 92
REFERENCES... 95
APPENDIX A: SURVEY QUESTIONS ... 100
APPENDIX B: SURVEY ANSWERS TABLE ... 107
APPENDIX C: AMR/AGV PROVIDERS WITH CASE STUDIES ... 108
Figure 1. Thesis process overview ... 2
Figure 2 Production network with AMRs (Fragapane et al. 2020 p.6) ... 10
Figure 3 Types of AMRs and application examples (Fragabone et al. 2021 p.407) ... 14
Figure 4 The design and evaluation framework for Operator solutions that support wellbeing (Kaasinen et al.2019 p.267) ... 24
Figure 5 Automation Acceptance Model (AAM) (Ghazizadeh 2012 p.45) ... 24
Figure 6 Central and decentral control architectures (De Ryck et al. 2020 p.153) ... 26
Figure 7 Matrix production (KUKA Ag 2016) ... 27
Figure 8 MiR100 (mobile-industrial-robots.com) ... 34
Figure 9 Layout in case study ... 36
Figure 10 Material flow in case study ... 36
Figure 11 Status lights of MiR1000 from MiR1000 user guide p.60 ... 38
Figure 12 AMR-AGV provider logos in manufacturing and logistics field (Logistics IQ 2021) ... 40
Figure 13 Literature review flowchart for refence cases ... 41
Figure 14 Localization map of MiR1000 ... 51
Figure 15 Damage on the factory floor ... 51
Figure 16 ROI in the first 25 months after the investment ... 56
Figure 17 Total distribution of the AMR/AGV reference cases ... 57
Figure 18 AMR reference case distribution in industry fields ... 60
Figure 19 Statistics of the sponsored Linkedin campaign ... 61
Figure 20 Distribution of the company sizes of respondents ... 62
Figure 21 Distribution of the annual sales of the respondents ... 62
Figure 22 Distribution of the respondent’s business type ... 62
Figure 23 Distribution of production type of the respondent’s companies ... 63
Figure 24 Importance of flexibility among respondents ... 63
Figure 25 Distribution of FMS usage in repondent’s companies ... 63
Figure 26 Distribution of the amount of FMS in use in respondents companies ... 64
Figure 27 Distribution of AMR familiarity among respondents ... 64
Figure 28 Distribution about the willingness of learn more about AMR suitability among respondents ... 64
Figure 29 Distribution of the reasons why the respondents are not interested of AMRs ... 65
Figure 30 Distribution of the AMRs in use in respondent’s companies ... 65
Figure 31 Distribution of different tasks of AMRs in respondent’s companies ... 65
Figure 32 Amount of production locations that the AMRs are serving in respondent’s companies ... 66
Figure 33 AMR equipment in use or repondet’s are interested in ... 66
Figure 34 Specifying answer for other question ... 66
Figure 35 Distribution of the satisfaction of AMR performance in the respondent companies ... 67
Figure 36 Distribution of the factors why AMRs were chosen in respondent companies ... 67
Figure 37 Specifying answer for the other question ... 67
Figure 38 Categorized deployment aspects from literature ... 69
Figure 39 ROI development after first 25 months from investment without fleet manager... 71
Figure 40 Industry type distribution of the production related reference cases ... 74
Figure 41 Industry type distribution for AMR cases ... 74
Figure 42 Significance and time budget chart for future research proposal ... 86
Figure 43 Example of AMR tasks in production ... 89
Table 1 Selected parameters for literature review ... 31
Table 2 Loading/Unloading timing ... 38
Table 3 Deployment aspects from literature ... 48
Table 4 Benefits of AMRs found from literature ... 49
Table 5 Disadvantages of AMRs found from literature ... 50
Table 6 Investment costs in case study ... 54
Table 7 ROI for the first five years after investment ... 55
Table 8 ROI for first 25 months in case study ... 56
Table 9 Reference cases related to production or manufacturing ... 58
Table 10 AMR/AGV reference cases from automotive industry ... 59
Table 11 Warehousing related reference cases ... 59
Table 12 Out of the box rreference cases for AMRs/AGVs ... 60
Table 13 Quantified deployment aspects from literature ... 69
Table 14 Quantified benefits from literature ... 70
Table 15 Quantified disadvantages from literature ... 70
Table 16 Reference cases dived in different industries ... 73
Table 17 Responses in which AMRs were already in usage ... 75
Table 18 Responses of companies that don't have AMRs yet ... 76
Table 19 Survey responses from batch manufacturing companies that are familiar with AMRs ... 77
Table 20 Names from companies shown in the logos but webpage was not found ... 79
Table 21 Survey answers from people not familiar with AMRs ... 81
Table 22 Comparison table for recent mobile robot related surveys ... 82
AGV Automated Guided Vehicle AI Artificial Intelligence
AMR Autonomous Mobile Robot
CNC Computer numerically controlled FIR Fourth Industrial Revolution FMS Flexible Manufacturing System
JIT Just-In-Time
PP Payback Period
RMS Reconfigurable manufacturing system
ROI Return on Investment
SBSM Small Bach Size Manufacturing
SLAM Simultaneous Localization and Mapping SME Small and Medium Enterprise
WIP Work In Progress
a acceleration/deceleration of AMR [m/s2] Cv Capacity of a vehicle [pcs/vehicle]
q productivity of each machine [pcs/h]
qAMR Theoretical throughput of each AMR [pcs/h]
L Length of the connecting path between production points [m]
M Number of production points
N Number of production lines
NAMR Total number of required AMRs [pcs]
Tc Theoretical cycle time [s]
tL/U loading/unloading time [s]
v average speed of AMR [m/s]
.
1. INTRODUCTION
The production batch sizes have been declining due to increased customer demand for greater product variety (Zandin 2001 C. 112.2.2). Small batch sizes (<10 pcs according to Zandin 2001 c.112.2.1), high demand fluctuation as well as random unexpected dis- turbances on the factory floor are characteristics of today’s production environment (Jä- rvenpää et al. 2016 p.120) Aside from productivity challenges, shorter production cycles and increased product customization requires a high degree of flexibility (Seder et al.
2019 p.95) and rapid reconfiguration (Järvenpää et al. 2016 p.120).
The new logistics concepts like autonomous mobile robots (AMRs) enable flexible and changeable production for a wide range of different products (Bozkurt et al.2020 p.38).
When 87% of the production time is going in transporting parts and components, the autonomous material handling has potential to make major impact on the production time in a typical factory (news.cion.com 2017). According to survey done by the Material han- dling institute (2021) 30% of the robotics and automation uses is the material and product movement within facility and the main barrier to adaption are the lack of business case and understanding of technology landscape.
This thesis is done for JTA Connection Oy. JTA provides tailored and turnkey solutions to boost production of an individual machine or the entire production line. JTA is also a leading supplier of innovative solutions for automatic material handling. However, AMRs are relatively new offering for the assigning company JTA Connection Oy. Based on AMR manufacturers sales arguments and a couple of earlier sales, JTA has a precon- ception that the AMRs are suitable for small batch size manufacturing (SBSM). To be able to offer AMR assisted production solutions to their customers, more information about the mobile robot capabilities, requirements and target markets needs to be gained.
The main goal of this thesis is to determine sales arguments for AMR assisted produc- tion. JTA can then focus their marketing to companies with suitable production assisting tasks for AMRs and justify the offers to those companies. Also improve their offer accu- racy with more knowledge about all considerable aspects. The approach to this main objective is done with three research questions:
- What aspects need to be considered when planning AMR assisted produc- tion?
- What are suitable tasks for AMRs in production environment?
- What is the current market state for AMR assisted production?
Due to the extensive nature of this this thesis, more than one research strategy is needed. The chosen research strategies are empirical, qualitative, and quantitative.
Based on these strategies four research methods were selected and those methods are:
- Literature overview - Case study
- Reference case review - Open survey
The data that is collected by using these methods is analyzed using suitable analysis methods and then combined to form partial results that are related to the main objective.
These partial results are deployment aspects, benefits and disadvantages, suitable tasks and current market state for AMRs in production usage. The thesis process is shown in Figure 1:
Figure 1. Thesis process overview
Deployment aspects are searched from AMR related standards and from other reliable literature sources. These aspects are tested and fulfilled in real customer case study and analyzed qualitatively. By using these same methods, a list of benefits and disad- vantages is formed. By searching real business cases from mobile robot suppliers and manufacturers web pages, the industry types are determined in which mobile robots have been used. After quantitative statistical analysis is performed, suitable tasks in pro- duction usage are presented. These tasks are then fulfilled with tasks that are merged from open survey results. Current market state is also determined by using the open
survey results that have been statistically analyzed and comparing results from different published surveys done recently.
This research concentrates on the small batch size production in Europe because JTA sees great potential in that market segment. Small batch size production domain has typically smaller level of automation and manufacturing tends to be produced in job shop mode with high labor content (Zandin 2001, c.112.2.1). Due to the thesis time frame, data were collected only from one case study and literature review is not done exten- sively.
The research is structured in seven chapters. After introduction, the theoretical aspects related to this thesis’s subject are presented in the second chapter to gain understanding of the research topic, the theory related to this research is presented in the second chap- ter. This chapter concentrates on the key elements of adaptive automation. Firstly, adap- tive automation paradigms are presented followed by information of production manage- ment. The third sub-chapter concentrates on cell type of layout. In the fourth sub-chapter autonomous mobile robots are introduced. After this, the deployment aspects found from literature are presented. The sixth chapter is about the benefits and disadvantages of AMRs in production usage. Then technology acceptance theories are explained. The last chapter is about the emerging fourth industrial revolution.
Followed by the theory chapter, the research process is presented thoroughly in the third chapter. In the first chapter, the research methods are explained, and selection justified.
Then research implementation of the different methods is described in detailed way in sub-chapters 3.2-3.5 each in their own chapter. The sixth sub-chapter is about the ex- plaining the selected data analysis methods as well as the analysis process. In the last sub-chapter, the used validity and reliability methodology are explained.
Research results are presented in the fourth chapter for all used methods separately.
After the results, the fifth chapter concentrates on the analysis and evaluation of these results. First analysis is performed for each result from different research methods. Then their reliability and validity are discussed in the second sub-chapter again one result at a time. The overall research process is evaluated in the sub-chapter 5.3. In the last sub- chapter the future research proposals are made.
Key findings are presented in the sixth chapter each sub-chapter presenting the summa- rized and combined results of each methods for all the research questions to give clear answers. Conclusions are then presented in the final main chapter.
2. KEY ELEMENTS OF ADAPTIVE AUTOMATION IN SMALL BATCH SIZE PRODUCTION
Manufacturing adaptability refers to the ability to respond to different manufacturing re- quirements and this includes components like flexibility and modularity (Zandin 2001 c.112.3.4). Aspects enabling automation to be adaptive to the dynamic marked changes starts from the production configuration (factory layout). The second level is the produc- tion management system and material flow. Then also the used automaton systems need to be adaptive. According to Zandin (2001), the production of smaller lot sizes re- quires flexible manufacturing systems to avoid time consuming chance from one product to another (Zandin 2001 c.112.3.4). Mobile robots are generally regarded as the material handling link providing the flexibility and responsiveness of flexible manufacturing sys- tem (Kumbhar et al. 2018 p.255).
Another viewpoint to this subject is the technology acceptance that is affecting the adop- tion of adaptive automation systems. Since the acceptance of employees can have a major impact on implementation of mobile robots (Markis et al. 2019 p.31), theory about technology acceptance is presented in the chapter 2.7.
Many researchers (Bortolini et al. 2018 p.103, Markis et al. 2019 p. 26, De Ryck et al.
2020 p.153) are stating that autonomous material handling solutions play crucial role for the future manufacturing. Therefore, in the last sub-chapter 2.8 visions for the future manufacturing are presented.
2.1 Adaptive automation paradigms
In the first sub-chapter, the flexible automation paradigm is explained. The second sub- chapter is about the reconfigurable manufacturing paradigm. Both paradigms are strongly linked to mobile robots (Kumbhar et al. 2018 p.255) and to the future of manu- facturing Bortolini et al. (2018 p.93,103).
2.1.1 Flexible manufacturing systems
System flexibility can be described as the ability to change to different states with low increase in time, cost, effort, or performance. A flexible system can assemble a variety of products using machines that are able to perform a number and variety of operations within pre-defined boundaries that allow change and adjustments (Svensson Harari et
al. 2018 p.549). Flexible manufacturing is empowered by versatile equipment and multi- skilled workers (Zandin 2001 c.112.3.4). For example, according to Koren (2006 p. 28) FMS often consist of computer numerically controlled (CNC) machines and other pro- grammable automation. However, for a system to be regarded as FMS, it does not need to be fully automated (Lenz 2016 c.38.4).
The flexibility of a manufacturing system has become as one of the key aspects for small batch size production (Lapinleimu et al. 1997 p.62) because production lot sizes tend to be small in flexible manufacturing (Zandin 2001 c.112.3.4). Lenz (2016 c.38.3.1) also states that FMS is an ideal solution for low inventory – high mix production since it is the least costly solution (Lenz 2016 c.38.5.2). Other benefits according to (Lenz 2016 c.38.5.2) are increased operating hours for machines simultaneously reducing the labor hours ratio for these machine hours. Also the quality is improved when parts are pro- cesses in planned paths and the movement monitored.
The downside of FMS is that to achieve high-capacity utilization, thorough plan and ex- ecution is required. Typical operation rate of an FMS varies from 55 to 65% Lenz (2016 c.38.3.1). According to De Ryck et al. (2020 p.153) there will always be trade-off between flexibility and optimality.
Lenz (2016 c.38.2.) states that any manufacturing operation with five or more part num- bers sharing the same process equipment qualifies as a flexible manufacturing system (FMS). When changing one part to another, no setup is required. This setup elimination is the primary characteristic that differentiates FMS from any production system.
Both internal and external factors can cause the need for flexibility (Burger et al. 2017 p.
37). Internal need for flexibility could be caused by machine breakdown (Čech et al. 2020 p.106). One example of the external need is the increasing need for product variants and customization (Zandin 2001 C. 112.2.2).
A literature review done by Svenson Harari et al. (2018) presented that there are dimen- sions, time horizon and elements for flexibility (Sevenson Harari et al. 2018 p.554). Also, many researchers have discussed different types of flexibility in their publications. Ac- cording to Bulger et al. (2017) four different flexibility types can have an impact for system level. Those types are variant spectrum flexibility, expansion flexibility, scheduling flexi- bility and volume flexibility (Burger et al. 2017 p. 35-36). Expansion flexibility refers to how easily the system can be expanded (Chyssolouris 2006 p. 23). Scheduling flexibility in FMS means that the system capacity can be adjusted to meet the demand. Also, the transportation paths within FMS can be allocated as needed (Lenz 2016 c.38.5.2.6). The path allocation can be considered as routing flexibility (Chyssolouris 2006 p. 23).
The process flexibility among SBSM is fundamentally different as among mass produc- tion industry. In SBSM, the process flexibility is complemented with product flexibility.
(Adli et al. 2020 p.1376-1377). Process flexibility means the ability to produce given batch by using different material or different ways. Product flexibility refers to how effi- ciently the production changed over from one set of products to another (Chyssolouris 2006 p. 23). Another type of flexibility type related to FMS is the layout flexibility and according to Kumbhar et al. (2018 p.255) it might be more important as per the applica- tion requirements
Flexible manufacturing systems have a wide range of applications according to Lenz (2016 c.38.4). Those applications can be for example welding, machining, assembly, or fabrication.
2.1.2 Reconfigurable manufacturing systems
Reconfigurable manufacturing system (RMS) is a recent class of manufacturing (Borto- lini 2018 p.94). Reconfigurability is an engineering technology that deals with design of production machines and manufacturing systems for cost effective and quick response to market changes (Koren, 2006, p. 27). RMS can be described as a machining system which can be created by incorporating basic modular hardware and software processes that can be rearranged or changed quickly and are simultaneously reliable (Mehrabi et al. 2000 p.404). Both hardware and software modules can be easily connected from the initial system design. An example of the hardware change can be the change of material handling system and software change can be re-programming the equipment. (Vavrik et al. 2020 p.1227).
The main objective for RMS according to Mehrabi et al. (2000 p.404) is to provide the needed functionality and capacity when it is needed. Reconfiguration allows flexibility in producing variety of parts as well as changing the actual manufacturing system itself.
The system modification can be done by removing, modifying, or adding certain process capabilities, controls, software, or machine structure. With these modifications, produc- tion capacity can be adjusted in response to changing marked demands or even tech- nologies (Mehrabi et al. 2000 p.404).
Koren (2006 p. 27) introduces six RMS core characteristics. Those characteristics are modularity, integrability, customization, scalability, convertibility and diagnosability. Vav- rik et al. (2020 p.1227) specifies that these characteristics can be implemented not only to the system level, but also to individual devices or to the company as a whole.
Reconfigurability can be divided into system level and machine level according to Koren (2006 p. 27).
Mehrabi et al. (2000 p.404-405) states that RMS is open-ended. This means that the system can be improved, upgraded, and reconfigured rather than the entire system re- placed. RMS goes beyond the economic objectives of FMS. This can be done by per- mitting minimizing lead time by reconfiguring existing systems or by adding new systems to the production. Also, rapid manufacturing modification and quick integration of new technology and/or new functions to the existing system are RMS objectives.
A typical RMS may include fleet of flexible and reconfigurable equipment like machine tools, robots and in-process inspection machines (Koren, 2006, p. 27). The workpiece material handling significantly influences the productivity and reliability of RMS. There- fore, RMSs are equipped with automated part handling and tool supply systems (Koren 2006 p. 36). In addition, the selection of the best configuration is among the most im- portant choices in the management of RMS (Bortolini et al. 2018 p.101).
According to Koren (2006 p. 28) the benefits of RMS include adjustable rates of produc- tivity and flexibility. Production can provide precisely the quantities needed at the lowest cost as possible. The greatest barrier towards the application of reconfigurable manu- facturing especially by the small and medium enterprises (SMEs) is the resistance to change (Bortilini 2018 p.94).
2.2 Production management
Production material management is explained in the first sub-chapter. The second chap- ter is about the material flow patterns.
2.2.1 Material management
Production planning process includes the planning of inbound and outbound values as well as the chosen sequence. Production planning makes production schedules for the next planning periods in advance and therefore directly impacts manufacturing control.
The key production planning tasks include planning the production requirements as well as the in-house production (Lödding 2013 p.85)
The production planning and scheduling becomes increasingly complex in reconfigura- ble environment since system must be reliable and quick at the same time (Bortolini et al. 2018 p.102). Processing of small batches requires detailed planning and control due to unique flow patterns and separation of workstations (Zandin 2001 c.112.2.3).
When planning manufacturing control, aspects like production planning and the structure of the product and production should be considered (Lödding 2013 p.539). Manufactur- ing control is influenced by following characteristics (Lödding 2013 p.99):
- Production strategy (Make-to-order, make-to-stock) - Spatial structure (cellular, job shop, etc.)
- Production types (mass, serial, batch, on-time) - Part flow
- Number of variants
- Fluctuations in customer demands
The number of produced variants is another highlighted production characteristic accord- ing to Lödding (2013). The manufacturing control tasks become more complex as the number of variants are increasing. Furthermore, the variant amount influences which production principle can be implemented. Also, the manufacturing process automation becomes more difficult. Variants should be formed as late in the production as possible.
This allows components to be manufactured beforehand and later customized for orders (Lödding 2013 p. 107-108).
For manufacturing control, it is critical that the load and capacity flexibility are sufficient as a whole to balance out the demand fluctuations. Capacity flexibility is the ability to quickly and efficiently adapt capacities to the changed capacity requirements. Load flex- ibility is the ability to adapt the load to the available capacities (Lödding 2013 p.110).
Most enterprises use already some of the available capacities during the production plan- ning to follow the expected demand fluctuations (Lödding 2013 p.479).
2.2.2 Material flow
The movement of the product through the facility is called production flow. The flow is influenced by the products, production environment, facility layout and the operational strategy (Zandin 2001 c.112.1). The products size, type, demand and expected lead times have an effect on production flow (Zandin 2001 c.112.2.1). The production envi- ronment refers to the production type, for example mass or batch production (Zandin 2001 c.112.2.2). The operational strategy includes order sequencing and scheduling as well as the degree of automation (Zandin 112.2.4).
On the other hand, Work In Progress (WIP) levels and throughput times of a production are strongly influenced by the used part flow type (Lödding 2013 p.102). The frequent
movement of small lots is characteristics of continuous production flow (Zandin 2001 c.112.3.2).
Lödding (2013) introduced three types of part flows and those types are lot-wise trans- portation, one-piece flow and overlapping production. In lot-wise transportation a com- plete lot is processed on a certain workstation and the transport to the next working sta- tion occurs only after processing the last part of the lot. The suitable throughput time in production is approximately four times the operation time of a lot (Lödding 2013 p.102).
Continuous one-piece flow represents the ideal transfer lot size and flow between work- ing areas. One-piece refers to a container size. In this flow type the number of pieces or work-in-process (WIP) should be established beforehand to avoid stopping the whole production plant for example due material shortages. To reach the ideal one-piece flow, minimizing the material flow between equipment and workstations needs to be achieved and often this requires layout changes (Santos et al. 2006 p.21). Ideal logistics case would be the direct part transportation to the following workstation (Lödding 2013 p.102).
Throughput time is a bit longer than the longest operation’s throughput time of an equiv- alent production with lot-wise transports (Lödding 2013 p.104).
Overlapped manufacturing refers to manufacturing in which an already processed part of a lot is transported to the following workstation and continued to be processed there.
Possibilities for implementing overlapped manufacturing are accelerating individual or- ders if needed and the other is reducing the attainable throughput times and WIP levels by distributing order for different working stations. Transportation requires additional ef- fort, but better utilization of workstations is gained. There are also fewer demands on the layout (Lödding 2013 p.105-106).
Material flow should be considered before buying new equipment to find the optimal lo- cation because the new equipment replacement might affect the material flow on larger or smaller scale (Santos et al. 2006 p.19-20). Categories that affect to the production flow are product, production environment, facility layout and operational strategy (Zandin 2001 c.112.1).
Smaller batch sizes increase the material handling between sections (Santos et al. 2006 p.21). Reduced batch size help to minimize the disadvantages of lot-wise productions.
Other advantages are shorter collection and preparation times and more constant load on downstream workstations (Lödding 2013 p.107). Buffers can be used to smooth out material flow variation. These buffers can be in two forms: material and time (Zandin
2001 c.112.3.6). AMR stations can be situated before and after working stations for buff- ering as seen in Figure 2. These actives can be supported by installing small conveyors on the AMRs (Fragapane et al. 2020 p.6).
Figure 2 Production network with AMRs (Fragapane et al. 2020 p.6)
The material flow complexity is influenced by two factors according to Lödding (2013).
Those factors are the number of workstation and the arrangement of those stations. Pro- duction line has the least complex material flow (Lödding 2013 p.108). Prudhvi et al.
(2018) stated that the layout design has a strategic role in successful establishment of flexible manufacturing system. In terms of flow patterns, they divide layouts in to three categories: conventional, single loop and tandem layout. Conventional layout is divided into unidirectional and bidirectional flow patterns. In unidirectional layout, the vehicle flow is only one direction and in bidirectional in two directions. In single loop vehicle flow occurs in a single loop. In tandem layout is based on zoning (Prudhvi et al. 2018 p.3982- 3983).
With AMRs the autonomous process flow among workstations can be improved and also new human robot interactions enabled (Klančar et al.2017 c.7.4). Mosallaeipour et al.
(2018 p.83) have studied mobile robots scheduling in their case study to achieve opti- mized cycle time in cell production. According to them, it is beneficial to use mobile robots to serve manufacturing cells, but the mobile robot scheduling becomes complex. The scheduling complexity comes mainly from the facts that the number of active machines varies as well as the production volume. Other aspects that increase complexity is the number of buffers because this increases the number of locations that the mobile robots need to serve.
2.3 Cell layout configuration
When batch sizes are decreasing and product variety increasing, a line type layout no longer becomes efficient. Also, the inefficiency that characterizes a job shop layout is not
cost effective I today’s competitive environment. Therefore, a cellular-type layout be- comes more effective (Zandin 2001 c.112.2.3). For manufacturing companies, the fac- tory layout depends on the product type and manufacturing volume. Primarily the layout is the result of materials flow in the production plant (Santos et al. 2006 p.25). Low in- ventory combined with high product mix is the most complex production configuration according to Lenz (2016 c.38.3.1). The layout design and optimization are key elements for reconfigurable manufacturing systems since different layout configurations are needed to change from one product family to another (Bortolini et al. 2018 p.100-101).
In cellular-type layout each family of parts have their own manufacturing cell. The product family is formed from products that utilizes similar machines, tooling and labor skills. After the cell formation, the cells are then organized within the factory space in a way that the best production flow is achieved. The most common configuration for production cells is a linear or U-shaped configuration (Zandin 112.2.3). Different production cells are linked with each other by automated material handling devices (Chyssolouris 2006 p. 23). To minimize the material handling cost, mobile robots can be used as these devices (Borto- lini et al. 2018 p.100).
According to Santos et al. (2006 p.47) following phases can be used for designing a production cell:
- Formation of product families - Change of the machine’s location
- Output production rate calculation, tasks assigning for each workstation and de- termine necessary number of workers in the cell
- Plan and control the cell
When employed properly, a cellular layout can enable the elimination of unnecessary product movement and both product and labor flexibility can be achieved. Also, more economic production environment can be achieved (Santos et al. 2006 p.44). According to Zandin (2001 112.2.3), the challenge is balancing each machines capacity and syn- chronize the production.
The dynamic production environment leads to the formation of dynamic cellular manu- facturing systems in the future. In those systems, cells are reconfigured within multiple planning periods. In Reconfigurable cellular manufacturing systems (RCMS), cells have modular machines. Those machines are composed by basic and auxiliary modules to be able to perform different operations and tasks (Bortolini et al. 2018 p.101).
2.4 Autonomous mobile robots
Logistics systems are one of the core components in the manufacturing industry (Me- hami et al. 2018 p.1077). Different terms are used when talking about mobile robots.
The most common terms are automated guided vehicles (AGVs) and autonomous mo- bile robots (AMRs) (Markis et al. 2019 p.26). Both AMRs and AGVs can be used inter- changeably to mean autonomous devices that can execute industrial tasks (Oyekanlu et al. 2020 p. 202313). To understand the similarities and differences between AGVs and AMRs, a short description of AGVs is given in the following paragraph.
AGVs are automatically controlled vehicles that are used in-house for floor bound con- veying (Markis et al.2019 p.26). AGVs have been widely used in automated material handling in industrial environment for over a decade according to De Ryck et al. (2020 p.152). The vehicle guidance is realized by some hardwire like wires or lines that are placed in the working environment of the robot (Markis et al. 2019 p.26). The conven- tionally used AGVs require also a well-defined, structured and obstacle-free working en- vironments according to Mohammed et al. (2019) and De Ryck et al. (2020 p.152,163).
The history of AMRs begins in 1987 when the first generic AMR patent was issued by Mattaboni (Mattaboni, 1987). Since then, the material handling technology has advanced rapidly and the need for more flexibility has driven the development of AMRs (Fragapane et al. 2021 p.405). In general, AMRs are industrial robots (Fragapane et al. 2021 p.407) that are an advanced form of AGVs (Oyekanlu et al. 2020 p. 202313), since AGV material handling system have evolved from laser guidance to today’s vision-based navigation of AMRs (Fragapane et al. 2021 p.405). In contrast to AGVs, AMRs can move around au- tonomously and perform a specific task (Markis et al.2019 p.26). AMRs use a decentral- ized decision-making process for collision free navigation. This kind of navigation pro- vides a platform for material handling, collaborative activities as well as full services within a bounded area (Fragapane et al. 2021 p.407).
The core tasks that are handled autonomously by the mobile robots are the localization, path planning, motion planning and vehicle management (De Ryck et al. 2020 p. 154).
The first three of these core tasks are related to navigation. AMRs navigation utilizes sensors, powerful on-board computers, artificial intelligence (AI), and simultaneous lo- calization and mapping (SLAM) technology (Fragapane et al. 2021 p.405). In SLAM tech- nique mobile robots gather information from their close surroundings and construct a map based on that information. By incrementally extending the map boundaries, the ro- bots can then determine their localization within that extended map (Kangan et al. 2020
c.6.4). The last core task of vehicle management includes the status monitoring of battery status, errors, and required maintenance (De Ryck et al. 2020 p.154).
AMRs operations can be controlled using human machine interface (HMI). By using a designed HMI, the behavior of the mobile robot fleet can be influenced and controlled (Oyekanlu et al. 2020 p. 202330). For example, an Android smartphone can be used as HMI (Oyekanlu et al. 2020 p. 202327). With the fleet manager in place, the mobile robots can communicate and avoid collision with each other. Also, the AMR performance can be improved by scheduling done by fleet management (Oyekanlu et al. 2020 p. 202328).
The main task of a fleet manager is to optimally assign a set of tasks to a set of mobile robots (De Ryck et al. 2020 p. 154).
According to Kangan et al. (2020), different types of ground moving mobile robots exist.
There are also different wheel types like Hilare-type and omni-wheeled vehicles. Omni- wheeled vehicles can move at any given configuration in all planar directions. (Kangan et al. 2020 c.3.1). More commonly used steering configuration for mobile robots is the Hilare-type also known as differential steering. A mobile platform robot which uses this type of steering has two independent motors that control the rotational velocity of each wheel (Kangan et al. 2020 c.3.2).
Autonomous mobile robot hardware components consist of various mechanical and elec- tronic parts. Mechanical parts are for example the robot body and wheels. Electrical parts are actuators, sensors, computers, power units and telecommunication related parts.
(Klančar et al. 2017 c.1.1.4).
The most typical applications for mobile robots are according to Patricio & Mendes (2020 p. 145) following:
- Transportation of raw materials from warehouse to production and finished prod- ucts back to warehouse
- Holding or buffering material over a period of time for storage
- Distribution tasks like package and bulk break handling as well as shelf fulfillment - Enabling Just-In-Time (JIT) assembly and reducing line-side storage require-
ments
AMRs can be divided to three different categories by the performed activities and char- acteristics. The categories are material handling, collaborative and full service. There are also three working environment categories: manufacturing, warehousing and other in- tralogistics environments. The AMR types can be seen on the Figure 3 (Fragapane et al.
2021 p.406). It is also a possibility for AMR to belong in multiple categories at the same
time. For example, Oyenkanlu et al. (2020 p.202330) have presented that AMR design can be based on hybrid between towing and loading.
Figure 3 Types of AMRs and application examples (Fragabone et al. 2021 p.407) A mobile manipulator is a system that combines a mobile robot and a serial manipulator equipped with sensors and at least one actuator. The actuator is usually a gripper system for object manipulation like part picking (Markis et al. 2019 p.27). Figures of mobile ma- nipulators can be seen in the previous Figure 3 in first two rows full-service column as
well as in the first row of the collaborative column. Oekanlu et al. (2020) suggest that in case of AMRs, the mobile base and the robot arm can be regarded as separate subsys- tems (Oyekanlu et al. 2020 p. 202313).
Currently hundreds of suppliers worldwide supply AMRs (Fragapane et al. 2021 p.406).
For example, a market study done by Logistics IQ (2021) resulted 275 AGV-AMR pro- viders just in manufacturing and logistics segment alone. Companies like Kuka Ag, Mo- bile Industrial Robotics (MiR) and Fetch robotics were mentioned.
2.5 Deployment aspects for autonomous mobile robots
Information about the things to consider when planning AMR implementation in produc- tion environment are regulated by different standards. In the first sub-chapter, all the related standards are mentioned, and key points of each standard related to deployment aspects are listed. The deployment aspects have also been discussed in other literature.
In the second sub-chapter, deployment aspects that can be found from the literature are described.
2.5.1 Deployment aspects from standards
According to Rose (2020) and Markin et al. (2019) due to rapidly developing capability of mobile robots, the safety of human workers around mobile robot applications needs to be ensured. Standard ISO 3691-4 is applicable for mobile robots as well as new Amer- ican standard R15.8 part 1 safety requirements for the industrial mobile robots. Prior these recent standards mobile robot manufacturers had only general safety require- ments for industrial machinery (Rose 2020 p.1, Markin et al. 2019 p.26). General stand- ards for machine safety are presented in ISO 13849-1.
A general standard for machine safety (ISO 13849-1 p.51-52) lists following aspects for machine usage:
- Usage restrictions including environmental conditions - control methods
- maintenance
Standard for Industrial trucks and their systems (ISO 3691-4 p.37) specifies the environ- mental conditions as:
- cleanliness and conditions of paths and floor markings
- freedom of paths from obstacles which can have impact on truck movement and clearance
- removal of spillage, dust etc. from paths to avoid skidding
Standard for Industrial trucks and their systems (ISO 3691-4 p.38) specifies that the manufacturer should give specifications for following floor condition characteristics:
- flatness - strength - surface finish - capping e.g. drains - metal content - underfloor services - electrical conductivity - joint position and quality
- permitted different level of floor e.g. slopes and steps
Standard for Industrial trucks and their systems (ISO 3691-4 p.40-44) adds that the man- ufacturer should provide following instructions when putting driverless trucks into to ser- vice:
- recommendation of permanent floor/ground marking for operating hazard zones for example the load transfer area
- operating zone should have minimum clearance of 0,5m wide for height of 2,1m on both sizes of the path for escape route
- the maximum allowable speed is dependent on the operation zone. The highest speed is 1,2 m/s
Haxhibeqiri et al. (2018 p.22) introduces the standard for wireless communication of mo- bile robots. The standard for wireless network of AMRs is IEEE 802.11. This standard gives information of for example the data transfer speed and band width. Section 802.11g is for places that wide band width is used. Industrial halls with fleet of AMRs are an example of that kind of place.
More standards are applicable for a top-module or equipment that can be used with mo- bile robots. For example, EN 1570-1 Safety of lifting tables is applicable for lifting table -
top module. Other example is that different standards should be applied for mobile ma- nipulators which are a combination of mobile platform (for example a mobile robot) and robotic arm, depending on the purpose of this robotic arm. However, there are currently no fully complied standards for this type of combination. Current standards are following (Markis et al. 2019 p. 28):
- ISO TS 15066:2016 Robotics and robotic devices – collaborative robots
- ISO EN 10218-1:2011 Robots and robotic devices – Safety requirements for in- dustrial robots
2.5.2 Deployment aspects from literature
Typically, in manufacturing environment, the number and type of tasks are higher com- pered to warehouse environment (Fragapane et al. 2021 p.414). It is therefore important to decide the tasks for AMRs at the early stage because this will have an impact to many other aspects (Čech et al. 2020 p.106). One of which is the number and type of vehicles (Fragapane et al. 2021 p.409). The needed number of mobile robots is mainly affected by the travel distances and production volume (Fragapane et al. 2021 p.411). To be able to calculate the accurate AMR fleet size, following aspects need to be determined (Čech et al. 2020 p.106 & 108):
- Supply need
- AMR load capacity in terms of how many units can be transported at a time - The average speed for AMR
- Manipulation times (loading and unloading) - Charging time
- Time reserve for other scheduled (e.g. maintenance) or unexpected events (e.g.
breakdowns)
Fragapane et al. (2020 p.7-9) have presented equations 1-3 for calculating the total num- ber of required AMRs. In these equations, the charging time and maintenance have not been taken to account. Firstly, the theoretical cycle time (Tc) in seconds is calculated with equation 1:
𝑇𝑐 =𝐿
𝑣+ 2𝑣
𝑎+ 2𝑡𝐿/𝑈 (1)
In which Tc is the theoretical cycle time [s], L is the length between two production points [m], v is the average speed of AMR [m/s], a is the acceleration/deacceleration of AMR
[m/s2] and tL/U is the loading/unloading time [s]. Then using the equation 2, the theoretical throughput (qAMR) of each AMR in terms of the number of transported products pers hour:
𝑞𝐴𝑀𝑅 =3600
𝑇𝑐 ∗ 𝐶𝑣 (2)
In which qAMR is the theoretical throughput of each AMR [pcs/h] and Cv is the capacity of each vehicle [pcs/vehicle]. Then the number of required AMRs (NAMR) is calculated using equation 3:
𝑁𝐴𝑀𝑅 = (𝑀 − 1) ∗ (𝑁∗𝑞
𝑞𝐴𝑀𝑅+ 4𝑁) (3)
In which NAMR is the total number of required AMRs [pcs], M is number of production phases [pcs], N is number of production lines and q is the productivity of each machine [pcs/h]. In this previous equation 3, the last term 4N means the assumption that at least two AMRs would always be available in front of each machine (Fragapane et al. 2020 p.9). If this is not desired, the equation can be simplified to following equation 4:
𝑁𝐴𝑀𝑅 = (𝑀 − 1) ∗ (𝑁∗𝑞
𝑞𝐴𝑀𝑅) (4) If the number of production phases M is only 1, the equations 3 and 4 would result 0 (1- 1=0). Therefore, the equation 4 can be further simplified to equation 5 to be able to cal- culate the number of AMRs for one phase production:
𝑁𝐴𝑀𝑅 = N∗q
𝑞𝐴𝑀𝑅 (5) Although the capacity calculation variables should be determined by the fullest produc- tion capacity, according to Čech et al. (2020) the irregular peaks in demand should not be covered by increasing the AMR fleet size. The larger number of AMRs often results decrease in the AMR efficiency and utilization rate (Čech et al. 2020 p.108). Fragapane et al. (2021) also clarify that it is possible use different types and size of AMRs and their functions can be different within the same fleet (Fragapane et al. 2021 p.411).
Čech et al. (2020) suggest the execution of cost-benefit analysis. In this analysis, the initial investments and operation costs are compared to the gained saving and return on investment. Comparison between selected vendors and technologies is also recom- mended (Čech et al. 2020 p.106). To evaluate AMR suitability from economical point of view (Fragapane et al. 2020 p.9-10). According to Philips & Philips (2019 c.1) the return on investment (ROI) can be calculated using following equation 6:
𝑅𝑂𝐼 (%) =𝑁𝑒𝑡 𝑃𝑟𝑜𝑔𝑟𝑎𝑚 𝐵𝑒𝑛𝑒𝑓𝑖𝑡𝑠
𝑃𝑟𝑜𝑔𝑟𝑎𝑚 𝑐𝑜𝑠𝑡𝑠 ∗ 100 (6)
According to Philips & Philips (2019 c.1) in order to ROI being truly meaningful, it should be reported with other performance measures. Such measure could be the payback pe- riod (PP), which is used to determine the time period after the investment when the in- vestors can expect to recover their investments. The break-even points also correspond the 0% of ROI. The equation for the PP is shown in the equation 7:
𝑃𝑃 = 𝑃𝑟𝑜𝑔𝑟𝑎𝑚 𝐶𝑜𝑠𝑡
𝑃𝑟𝑜𝑔𝑟𝑎𝑚 𝑏𝑒𝑛𝑒𝑓𝑖𝑡𝑠 (7) Zandin (2001) bring up the aspect of how the item is fed to the machine (Zandin 2001 c.108.4.7). In this case, how the items are given to the AMR for transportation and also how the items are moved or handled after the transportation. The AMR operation possi- bilities can be extended by adding manipulation equipment and those should be planned for short and long term (Fragapane et al. 2021 p.409). Čech et al. (2020) concluded in their research that AMRs usually require customization by special top modules to serve production. The needed top modules could be for example lift, hook or conveyor (Čech et al. 2020 p.106-107). Another additional technology that might be required is the iden- tification of the transported parts. This can be accomplished for example by adding a camera for QR-code recognition (Čech et al. 2020 p.106-107). It is also possible to im- plement various type of equipment (Fragapane et al. 2021 p.411) and even sharing equipment within the same AMR fleet (Fragapane et al. 2021 p.422). Other aspect to achieve autonomous loading and unloading is the possible need of docks or racks.
These kind of floor fixed accessories increase the space requirements (Čech et al. 2020 p.106-107).
Haxhibeqiri et al. (2018 p.19) explain the need for wireless communication solution. Due to the fact that factory environment is often a large area, multiaccess point (AP) systems should be used to cover the entire factory area. The communication between mobile robot and fleet management controller requires low network latency.
The overall system responsiveness and efficiency can be improved by zoning and ser- vice point configuration (Fragapane et al. 2021 p.413). In zoning, the factory entire fac- tory floor area is dived in zones and workflow of each zone are serviced by one or more mobile robot (Oyenkanlu et al. 2020 p.202330, Fragapane et al. 2021 p.413). By distrib- uting the workload in zones, the travel distance between service points can be reduced and traffic and throughput time minimized. These improvements can be achieved by an- alyzing the operation area and configuring the needed service points (dynamic or fixed) and delimiting and replacing zones for each mobile robot (Fragapane et al. 2021 p.413).
One robot is often capable of serving multiple machines (Zandin 2001 c.108.4.7).
Another fleet optimization method is scheduling mobile robots simultaneously with other machines, equipment, and humans. This can be done by using mathematical and simu- lation models. The benefits of scheduling are reduced waiting times and transportation costs (Fragapane et al. 2021 p.414). Smart dispatching methods are also related to scheduling. An example of smart dispatching is that the mobile robots could be close to demand point before the actual need occurs (Fragapane et al. 2021 p.415).
Markis et al. (2019 p. 29) bring up the fact that in addition to human safety, the safety of mobile robot itself as well as the transported products should be considered especially in industrial environments to avoid economical damage.
Quadrini et al. (2020) bring up the aspect that automated logistics systems like mobile robots often need to be integrated with existing production managements systems.
These systems can be for example Manufacturing execution System (MES) and Enter- prise Resource Planning (ERP) (Quadrini et al. 2020 p.407). When defining the inter- faces between AMRs and their operating environment, it is crucial to determining which parts of the system should be controlled in centralized or decentralized manner accord- ing to (Fragapane et al. 2021 p.410). In facilities with small AMR system size, more cen- tral control is referable, and decisions can made by a cloud-based system. With cloud- based control system, the overall costs can be kept lower (Fragapane et al. 2021 p.410- 411).
2.6 Benefits and disadvantages of AMRs in production envi- ronment
There are many aspects that make the use of mobile robots appealing (Klančar et al.
2017 c.1.1.4). According to Zandin (2001), the basic objective regarding automation adoption is to improve the operator’s work environment by reducing dangerous and harsh work, simple and boring operations as well as unhealthy working conditions. Au- tomation can also be used for reducing operator headcount, personal errors, improve productivity, balance cycle time and enhance nighttime utilization rate (Zandin 2001.
c.108.4.1).
Flexibility and adaptability of AMRs are often the most endorsed benefit by researchers and by AMR manufacturers. Mobile robots are flexible in their use (Markis et al. 2019 p.26) and can adapt quickly to changes in the operating environment (Fragapane et al.
2021 p.405) due to situation aware movement planning (Markis et al. 2019 p.26). The strength of AMRs is well demonstrated in narrow-aisle, high-traffic environments that require dynamic motion planning (Fragapane et al. 2021 p.406). Karabegovic et al. (2021 p.193) have stated that AMRs have 98,5% ability to work itself. AMRs are even able to
use elevators to move between different floors as seen in the YLE news (YLE 2021).
The intelligence of AMR system also allows them to be close to the point of demand before and actual need is announced (Fragapane et al. 2021 p.407).
Scalability is also another highly endorsed benefit. According to Fragapane et al. (2021), a fleet of AMRs can be easily scaled for increasing or decreasing the number of AMRs in use. This change in the fleet size does not require any structural change (Fragapane et al. 2021 p.407). AMRs also allow dynamic customer specific modifications that can be implemented rapidly (Markis et al. 2019 p.26). Fragapane et al. (2021 p.407) also bring up the robustness of AMRs as a benefit. The robustness is related to the autonomous decision making and therefore AMRs can recover after failure or mission interruption.
Increased safety is brought up by Čech et al. (2020 p.107) and Klančar et al. (2017 c.7.4) in their research. Collaborative operations are regarded as one of the AMR benefits as well (Markis et al. 2019 p.26, Fragabane et al. 2021 p.407). AMRs are able to work to- gether with humans and also with other AMRs (Fragabane et al. 2021 p.407). Klančar et al. (2017 c.7.4) also state the improvement of workers health as an advantage.
The cost related benefits are often used in AMR marketing. Fragapane et al. (2021 p.407) highlights the cost efficiency as one of the benefits of AMRs in their research as well. According to Klančar et al. (2017), automated material handling reduces costs that are related to labor, productivity, inventory, and space. Also, throughput is improved (Klančar et al.2017 c.7.4). AMRs are able to operate 24/7 and have 15 years operation life (Karabegovic et al. 2021 p.193). With these savings in many areas, Karabegovic et al. (2021 p.193) states that investment on AMRs will pay off in 2-3 years.
The ease of integration to the entire factory production architecture is beneficial as well.
AMRs can be integrated quickly into a factory or other facility (Fragapane et al. 2021 p.407, Oyenkanlu et al. 2020 p.202331). Oyenkanlu et al. (2020 p.202331) specifies that, the integration does not need any supporting infrastructure like wires, optical markers or magnets for quidance. And as stated before in this chapter, later modifications of the AMR system do not require structural changes to the existing production environment.
Oyenkanlu et al. (2020 p.202335) also continues that with the AMR integration to the production and warehouse databases, production material tracking can be improved.
Future benefit is according to Fragapane et al. (2021 p.407) and De Ryck et al. (2020 p.152) the decentralized control with intelligent, cognitive, and behavior-based method- ologies and technologies. With the decentralized decision making, the flexibility and productivity can be maximized.
