Available online at www.sciencedirect.

com

ScienceDirect
Procedia Manufacturing 9 (2017) 81 – 88

7th Conference on Learning Factories, CLF 2017

Learning in the AutFab – the fully automated Industrie 4.0 learning

factory of the University of Applied Sciences Darmstadt
Stephan Simonsa*, Patrick Abéa, Stephan Neserb
a
Hochschule Darmstadt, Fachbereich Elektrotechnik und Informationstechnik, Birkenweg 8, 64295 Darmstadt
b
Hochschule Darmstadt, Fachbereich Mathematik und Naturwissenschaften, Schöfferstraße 3, 64295 Darmstadt

Abstract

The manufacturing industry is currently changing from mass production to customized production, which results in
challenges concerning different production aspects. Industrie 4.0, a German strategy, combines several technologies,
to meet these challenges. This implies that students need to acquire new competences. Therefore the University of
Applied Sciences Darmstadt has established a holistic, fully automated learning factory covering most areas of
Industry 4.0. Students of different degree courses learn very successfully in projects as well as in lab courses taught
in this smart factory. This paper describes the challenges for production and the technologies proposed by Industrie
4.0 to meet these challenges. It presents the fully automated Industrie 4.0 learning factory and the education within
this production facility as problem-based lab work and project-based courses.
©©2017
2016Published
The Authors. Published
by Elsevier B.V. by
ThisElsevier B.V.
is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the scientific committee of the 7th Conference on Learning Factories.
Peer review under responsibility of the scientific committee of the 7th Conference on Learning Factories
Keywords: Industry 4.0; IoT; Learning Factory; Smart Factory; Digital Transformation; Digitalization; Project-Based Learning; Problem-based
learning; Manufacturing Engineering; Production Planning and Control

1. Introduction

The manufacturing industry faces the challenge of switching from mass production to customized production with
a high production cost pressure at the same time. The productivity, the product quality and the flexibility of the
production have to be increased, while the delivery times and the inventory must be reduced. In this context,
flexibility means both the flexibility for different production quantities as well as the flexibility to produce different

* Corresponding author. Tel.: +49-6151-16-38314; fax: +49-6151-16-38931.

E-mail address: stephan.simons@h-da.de

2351-9789 © 2017 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer review under responsibility of the scientific committee of the 7th Conference on Learning Factories
doi:10.1016/j.promfg.2017.04.023
82 Stephan Simons et al. / Procedia Manufacturing 9 (2017) 81 – 88

types or differently configured products with the same production system without or with only very short setup
times. This also includes the flexibility to quickly switch productions for different customer orders from one
production unit to another, for example either when a unit fails or to reduce delivery times. Additionally, the time to
market of new products has to be shortened, which results, amongst others, in shorter times for the installation of
production facilities. At the same time, the complexity of production facilities increases in consequence of an
increasing intelligence of the products. To meet the growing complexity of work, workers should be assisted by
electronic systems to improve their efficiency. Furthermore, the engineering costs for the development of the
production units shall be reduced, for instance, by automatic engineering or by synchronous engineering of the
mechanical, the electrical and the automation department throughout the total lifecycle of the product and the
production facility.
Since plants of global companies are located worldwide, there is also a need for a secure remote access to the
plants in order to enable a remote diagnosis of failures as well as the possibility to update the control software
remotely. Connecting a plant to the Internet provides the opportunity to use services and storage in the cloud,
thereby reducing or scaling the costs to the actual needs. These include big data analystics for preventive
maintenance or web shops, in which the customer can configure the product by himself during the ordering process.
In addition, to all these aspects, the energy efficiency of the plant shall be improved.

2. Industrie 4.0 technology

In order to meet all these challenges the strategic initiative Industrie 4.0 was launched in Germany. Industrie 4.0
is part of the Federal Government’s high-tech strategy [1]. The key technologies to meet these challenges are an
increase of automation technology in the production plants and a seamless integration of IT systems and automation
technology.
According to [2], the five central paradigms of Industrie 4.0 are a vertical and horizontal integration, a
decentralized intelligence, a decentralized control, a fully integrated digital engineering and a cyber-physical
production system. The vertical and horizontal integration is implemented by connecting the production systems
with each other, so that there is a seamless data flow between all systems and even to the Internet. Decentralized
intelligence describes the ability to deliver relevant production information to decentralized control systems. The
fully digital engineering makes it possible to simulate the production system using digital twins. Cyber-physical
production systems (CPPS) consist of cyber-physical systems (CPS) which are embedded systems that use sensors
and actors to interact with the physical word. They are connected to each other thereby can interact also with the
digital world, use data and services worldwide, and have multimodal human machine interfaces [3]. Very often it is
assumed that they are consumer goods. However, production systems must operate reliable for decades. At the same
time, new programmable logic controllers (PLCs), the classic system used in production, are introduced to the
market with additional features for Industrie 4.0, making it attractive to use a hybrid approach of PLCs and
embedded systems [4].
The paradigms of Industrie 4.0 are implemented with different technologies. The customized production with
batch size one can be implemented by using radio frequency identification systems (RFID) or data matrix codes
(DMCs) in order to identify the product individually. By this each workstation can perform exactly the task required
for the actual order. The production-relevant data from the identification systems and other intelligent sensors and
actuators can be passed on to decentralized control systems of the work stations, typically PLCs, or directly into a
connected cloud. The decentralized intelligence increases the flexibility of the production with regard to the type or
configurations of products, which is supported by self-configuring work stations. The decentralized control systems
also enable a self-learning production, which automatically adjusts its parameters if the quality deteriorates.
The fully horizontal and vertical integration makes it possible to record runtime data as well as quality data both
locally or in the private or the external cloud. This allows to optimize the production and to avoid unexpected,
operational downtimes by means of predictive maintenance, which increases the availability of the plant.
Furthermore, the complete networking enables automatic documentation, for example of scrap reasons, which could
reduce the time for changes when quality issues occur. Moreover, the assembled parts can be tracked automatically
to reduce the number of units in an eventual callback and the inspection results can be tracked as a proof of the
 Stephan Simons et al. / Procedia Manufacturing 9 (2017) 81 – 88 83

product quality. In addition, the vertical integration increases the flexibility, by directly connecting the business
processes with the automation processes, i.e. by connecting a manufacturing execution (MES) and an enterprise
resource planning (ERP) system directly with the line. Thus, amongst others, it is possible to reduce the delivery
times by choosing always the optimal production unit for the order.
Additionally, integrating the complete functional safety technology can reduce the time for the diagnosis if a
safety issue occurs. This can be supported by interactive visualization systems, not only on touch panels at fixed
locations, but also on industrial mobile devices or on consumer devices like tablet computers and smartphones. By
linking embedded cyber-physical production systems, the production facility can communicate with an operator in
case of a warning, for example by email. Furthermore, operators in the facility can be assisted in their work by
information from the Internet, which could be shown on smart glasses. A remote control of the plant enables
worldwide diagnostics and software updates. In addition, web-services can be used to optimize the plants or for new
business models, often using big data analytics. However, connecting the plant to the Internet, typically by using
web technologies, must be done in a secure way so that the company data are secure.
Within the fully digital engineering, 2D and 3D simulations can be created for virtual commissioning and for
simulation of cycle times, energy consumption or ergonomic aspects of a production facility. Virtual commissioning
can not only shorten the down times during commissioning and changes but can also reduce the production failures
during the start-up phase. Further information on these technologies, their classification and use cases can be found
in [5], [6] and [7].

3. Industry 4.0 and the challenges for engineers and universities

Concerning all the technologies of Industrie 4.0, it is necessary that engineers need to improve their professional,
social, methodical and personal competences [8] shown in figure 1. They need an interdisciplinary understanding of
systems, production processes, automation technology, information technology, ergonomic principles and of
business processes. In addition, the skills for cooperation and communication in interdisciplinary groups are of
crucial importance. The modern engineer in Industrie 4.0 must have a holistic perspective on complex production
systems.

PLCs Holistical system

Productivity
Embedded Systems comprehension
Communication Problem solving
Flexibility
technol. / Interfaces abilities
Visualization &
Quality Independency
Required competences of engineers

Assistance Systems
Crucial aspects of production

Availability Sensors, Actuators, Interdisciplinary

Reliability Identificat. Systems expertise
Crucial technologies

Knowledge of
Delivery time Security & Safety
production processes
MES, ERP, Knowledge of crucial
Inventory Assisted Production
PLM / PDM technologies
Knowledge of
Comissioning time Simulation Systems
business processes

Assisted Production Cloud services Project managment

Communication /
Remote Control Web technologies
presentation skills

Documentation Big data analytics Ability to abstract

Fig. 1. Crucial aspects of production, crucial technologies, required competences (following [9])
84 Stephan Simons et al. / Procedia Manufacturing 9 (2017) 81 – 88

For the universities this means to overcome the boundaries between established disciplines. Lectures should
enable students to independently solve complex problems. This can be achieved by problem-based or project-based
learning in learning factories.

4. The Industrie 4.0 learning factory AutFab of the University of Applied Sciences Darmstadt

During the last years the faculty of Electrical Engineering and Information Technology of the University of
Applied Sciences Darmstadt has established, together with the faculty of Mathematics and Natural Sciences, a
holistic fully automated Industrie 4.0 learning factory, called AutFab, covering most aspects of Industry 4.0 (see
figure 2). The AutFab has a size of 50m2 and is equipped with state-of-the-art production components used in
industry. The line assembles configurable relays with batch size one, by using radio frequency identification RFID
technology. These relays may have 2, 3 or 4 contacts, protection modules with a red or a green LED and different
protection technologies. Thus it is possible to research on the fully automated assembly of customer configurable
products using a one piece production flow.
The AutFab consists of a high bay storage, two assembly and two inspection stations. The high bay storage works
with a three axis motion controlled robot. A six axis industrial robot and a pneumatic press are used as assembly
stations. The first inspection station works with an automatic optical inspection reading all the two dimensional data
matrix codes on the components and detecting if all and the correct components are assembled. Additionally, the
weight of the product is controlled there. The second inspection station performs an electrical test which, using one
of the DMCs, adapts automatically to the product. Intelligent shuttles transport the components between the different
workstations.
The different workstations are controlled by modern safety programmable logic controllers, which are connected
via open process communication unified architecture (OPC UA) technology to other systems. All controllers are
networked using modern real-time capable Industrial Ethernet systems and in parallel also classical fieldbus systems
or the standard Internet communication protocols TCP/IP. The functional safety components are fully integrated (see
figure 3). Thus it is possible to analyze every failure from most of the human machine interfaces (HMIs).
The HMIs are implemented as interactive visualizations on the fixed touch panels of the workstations, on a big
central touch panel and on an industrial mobile operation panel using an industrial visualization system.
Additionally, HMIs are implemented on consumer mobile devices, i.e. smartphones and tablet computers using self-
written Android, Java and JavaScript applications, which are connected via a central OPC UA server to the work
stations and via a native OPC UA server to the robot controller. Another HMI has been implemented by user
application webpages on the webserver of the main PLC. Thus every personal or tablet computer in the network of
the AutFab is able to use this responsive visualization with its web browser.

High-Bay
Optical & Weight Electrical Storage
Inspection Inspection

Six-axis
Robot Shuttle
Pneumatic System
Press

Fig. 2. The fully automated Industrie 4.0 learning factory AutFab of the University of Applied Sciences Darmstadt
 Stephan Simons et al. / Procedia Manufacturing 9 (2017) 81 – 88 85

Production can be controlled by an MES, which is connected via OPC UA to the production system, enabling the
MES to download orders to the automation system and to get back tracking data, concerning runtime as well as
quality data.
The whole smart factory is simulated as a digital twin in 3D for virtual commissioning. This includes both the
3D-CAD-representation as well as the behavior model of the line. The simulation is connected directly to the six
workstations via OPC classic to test the control software. The material flow and the energy consumption are
simulated in 2D and in 3D in another tool. In parallel the total energy consumption of the whole production line is
measured with smart meters connected to the network and by this visualized on the HMIs. Additionally, the line can
be accessed remotely using modern security technologies. Totally more than 95 PLCs, PCs and embedded cyber-
physical production systems are networked within the AutFab. Using bus analyzer, it is possible to analyze the
systems connected via industrial Ethernet with web based frontends.
Research on assisting workers at the manual rework place using mixed reality on smart glasses - the combination
of virtual and real world information within glasses - is in progress. Additionally, realizing the connection of the
plant via OPC UA to an enterprise resource planning system (SAP ERP/ME) running in the cloud, is in progress.
Orders can be placed either by the HMIs of a central work station, the tablet computers or smartphones, by the
MES system or by the ERP system. This implements a fully horizontal and vertical communication, a precondition
for Industrie 4.0. Thus the Internet as well as complex IT-Systems like ERP, MES, 2D- and 3D-simulation systems
are connected directly to the automation system granting an online access.
All software tools are industrial used systems. The smart factory AutFab is designed modular so that it can be
continuously expanded with new technology for process planning and control. It can be used for teaching, for
research and for development.

Fig. 3. The core workstations of the AutFab with the fully integrated safety technology

5. Education in the AutFab

Concerning the differentiation of learning factories defined by the CIPR Collaborative Working Group on
„Learning Factories for future-oriented research and education“ [10], the AutFab, assembling real physical products
with a real value chain and with learning taking place on the factory-site, is a learning factory in the narrow sense.
86 Stephan Simons et al. / Procedia Manufacturing 9 (2017) 81 – 88

However, the training in the AutFab is actually not concentrating only on using the technology, but more on
implementing additional functionality with existing or new technologies.
Students of different degree courses are educated in the AutFab using two different teaching methods. They are
working in labs, realized either as standard labs or as problem-based labs, or they are working on projects in the
AutFab. The different methods are used because of the different course programs.

5.1. Lab Education in the AutFab

Students in the major Automation of the international Master of Science of Electrical Engineering program have a
standard lab within the lecture Information Systems in Industrial Automation. During this lab of 3 hours the students
establish the communication between the software system MATLAB/Simulink and the AutFab via OPC and analyze
the data they get.
Students of the major Automation of the bachelor program Mechatronics as well as of the bachelor program
Electrical Engineering work within the lab Visualization, a problem-based lab, in the AutFab. Problem-based
learning was first implemented at McMaster University Medical School in the 1960s in the field of medical
education [11]. It is defined as a student-centered pedagogy in which students work on real-world problems of their
future working field. The students shall “learn by doing” and the role of the teacher is a “guide-on-the-side”.
Problem-based learning emphasizes on multidisciplinarity, collaboration and group work which shall result in a high
quality end-product [12].
In the problem-based learning lab “Visualization” the students work on 5 dates for 3 hours each in the AutFab.
After having been introduced to the used industrial visualization system during the first lab, by working through a
practical example, always two students are dedicated to a workstation and get the specification for the HMI of this
station. The students then have to prepare an ergonomic design for their specific HMI based on the requirements of
the specification. After a review of this design with the supervisor, they implement it during the next four lab dates.
The focus in this lab phase lies on the independent work of the students. They have to understand their workstation
holistically und thus acquire also knowledge in other fields. The supervisors stand by for questions only. The result
of each group is graded after the last date, which is part of the total grade of the lecture. Up to know already 120
students took part in this lab.

5.2. Project-based learning in the AutFab

Students from four different bachelor programs, a German and an international master course from three different
faculties are taught in semester projects in the AutFab. Most of them come from the programs Mechatronics of the
faculty Mechanical Engineering and Plastics Technology and from the program Electrical Engineering of the faculty
of Electrical Engineering and Information Technology. Further students are in the program Optoelectronic and
Images Processing of the faculty Mathematics and Natural Sciences and some are from Economic Engineering, a
common program of the three faculties Economics, Electrical Engineering and Mechanical Engineering. The given
problems demand to acquire an interdisciplinary knowledge. Additionally, the different origins of the groups require
an interdisciplinary collaboration to fulfill at least some of the tasks. The students gather knowledge about all
technologies used for Industrie 4.0 in the AutFab.
The semester projects are put into practice by using the project-based learning method. They are based on the
same key features as problem-based learning – learning by doing on real-world problems, teachers being the “guide-
on-side”, multidisciplinarity, collaboration and group work as well as a high quality end-product of the project –
however they are also focusing on project planning and project managing as well as on presentation and
documentation.
Figure 4 shows the process within this course. At the beginning of the semester the teachers present different
project tasks to the students. The students independently form teams and choose one of the projects. Based on the
specification given by the teacher they generate a project plan with time scheduling, duties and responsibilities.
During the semester the students work independently on the project. Within weekly project meetings they discuss
problems and they may ask the supervisors for guiding assistance. The projects include different sub-tasks, for
example getting a deeper understanding of the given problem, problem analysis, understanding the AutFab, risk
analysis, research of information, contacting suppliers, design of solutions, selecting components, implementing
 Stephan Simons et al. / Procedia Manufacturing 9 (2017) 81 – 88 87

components in hard- und software, verification and validation of the solution. Off course each project will only cover
a limited amount of the complexity of the total learning factory and the students will mainly acquire technical
competence in this area. However they also learn about the other areas during the final presentations and they
improve their skills to enter related technological fields during the project. The amount of guidance needed, depends
on the complexity of each task but is limited as the students can get most of the information from the documentation
of preceding groups or from manufactures. The results of the work have to fit to the rest of the AutFab, which off
course limits the solution space. Another limitation is the maximum duration for a project of one semester which
also has major impacts on the time schedules of the projects. However the projects in the past have shown that these
limitations are not significant and that the students still have enough possibility to implement their own ideas.
At the end of the projects, the students provide a documentation of their work for a subsequent group or for
maintenance reasons and a poster for a general readership. For some projects the students generate a video instead of
a written documentation, like the Video “Tutorial Plant-Simulation” which is a tutorial video for one of the software
systems used for the simulation of the material flow and the energy consumption of the AutFab. These videos are
uploaded on the YouTube channel CRAatHDA of the Center of Robotics and Automation of the faculty. Thereby
tutorial videos for subsequent groups and for others interested are generated for a simple training in this subject.
Additionally, the students present the results of their project to the other project groups and the supervisors as a
lecture and with a practical demonstration. The talk is recorded with a camera. The students do not only get a
feedback from their technical supervisor, but, based on the video, also an additional feedback concerning their
presentation skills by a trained non-technical member of the trainer team of the University of Applied Sciences
Darmstadt. A video, in which a project team was accompanied during their project throughout the semester, was
uploaded on the CRAatHDA channel to show students prior to the start of the semester how this course is carried out
and to show potential employers how the students are trained within the AutFab during these projects.
Finally the projects are graded by the supervisors. The grade is based on the project management, the work during
the project, the final results, the documentation, the poster or the video and the presentation. Up to now more than
180 students have carried out a project on the AutFab. One of the results is the fully integrated functional safety
technology shown in figure 3, which was implemented by 26 students within 6 semester projects and 2 final theses.

Choice of Project Project Verification

Implemen- Verification Document.
Design Presentation Presentation
Grading
project planning realization Validation
tation Validation Poster/Video

Presentation
Guide by side Grading
project ideas

Fig. 4. Process flow within the semester projects: student process steps on top, teachers process steps underneath

6. Actual Research projects in the AutFab

The actual research projects in the AutFab include simulations for virtual commissioning using a modern physics-
based simulation system, directly connecting the AutFab to an SAP ERP/ME system running in the cloud and the
integration of a smart glass with mixed reality to assist operators and engineers. These projects are also supported by
students.

7. Conclusion

The manufacturing industry is currently changing from mass production to customized production, which results
in challenges concerning different production aspects, for example productivity, flexibility and quality. The German
initiative Industrie 4.0 combines several technologies, to meet these challenges for which students need to acquire
new competences. Therefore students at the University of Applied Sciences Darmstadt are educated within the fully
automated Industrie 4.0 learning factory AutFab in semester projects as well as in problem-based lab courses.
The continuous evaluation of the projects and the labs in the AutFab always showed very good results. The
students emphasize the practical relevance, the possibility to work independently, the assumed relevance for their
88 Stephan Simons et al. / Procedia Manufacturing 9 (2017) 81 – 88

later work and the good assistance by the teaching staff. They asked for more projects and in earlier phases of their
degree course. At the end of the theses, which at the University of Applied Sciences in Darmstadt are typically
carried out in industry, the majority of the industry supervisors positively reported the practical experience of the
students. This shows that the project-based learning in the AutFab has a sustainable positive effect in the student
education for these programs.
By now, seventeen Video dealing with projects in the AutFab have been uploaded on the YouTube channel
CRAatHDA. These videos already have more than 30.000 clicks showing the tremendous interest in this type of
knowledge transfer. Nowadays videos are picking up students in their daily lives.
The presented learning factory AutFab is also used as an Industrie 4.0 demonstrator to attract new students. The
direct presentation of electrical engineering to school groups and the presentation of the practical education to
company representatives encounter great interest. In addition, the operation and expansion of the learning factory
enables the teachers to continually expand their knowledge, which is then incorporated into the lectures.
The learning factory AutFab has proved successfully in practical education. It allows the investigation of various
complex questions in many areas of production management and production engineering. Students get acquainted
with all technologies, ideas, challenges, advantages and disadvantages of Industrie 4.0. They can link the theoretical
knowledge they have acquired during their studies by practically working on real problems. Additionally, they
improve their skills in project planning and management, their ability to independently learn the ropes of technical
unknown subjects and to solve complex problems with a structural approach, which also deepens their holistic
understanding of complex systems. They independently acquire missing knowledge from different sources. Working
in teams and contacting, for example suppliers improves their communication skills. The projects in the AutFab
increase the interdisciplinary knowledge of the students as well as their presentation and documentation skills.
The project labs carried out so far show that the students working in the learning factory are extremely motivated.
They are “infected” by the fascination of the used technology. The students learn a big variety of key technologies
and methods for their later work. They have the opportunity to get to know the future topics associated with the
subject of Industrie 4.0 in real applications with industrially used components and state-of-the-art methods in order
to contribute to the competitiveness of their future employers with this knowledge.

References

[1] Bundesministerium für Bildung und Forschung / Federal Ministry of Education and Research (BMBF): The new High-Tech Strategy,
Innovations for Germany, Berlin 2013 on http://www.hightech-strategie.de/de/The-new-High-Tech-Strategy-390.php, last visited 4.12.2016
[2] D. Siepmann: Industrie 4.0 Grundlagen und Gesamtzusammenhang, in A. Roth pub.: Einführung von Industrie 4.0 – Grundlagen,
Vorgehensmodelle und Use Cases aus der Praxis, Springer-Verlag, Berlin Heidelberg, 2016
[3] H. Kagermann, W. Wahlster, J. Helbig: Recommendations for implementing the strategic initiative INDUSTRIE 4.0: Final report of the
Industrie 4.0 Working Group, 2013
[4] B. Vogel-Heuser: Herausforderungen und Anforderungen aus Sicht der IT und der Automatisierungstechnik, in T. Bauernhansl, M. ten
Hompel, B. Vogel-Heuser: Industrie 4.0 in Produktion, Automatisierung und Logistik, Springer Vieweg, Wiesbaden, pp. 37-48, 2016
[5] T. Bauernhansl, M. ten Hompel, B. Vogel-Heuser: Industrie 4.0 in Produktion, Automatisierung und Logistik, Springer Vieweg, Wiesbaden
2016
[6] W. Huber: Industrie 4.0 in der Automobilproduktion, Springer Vieweg, Wiesbaden, 2016
[7] P. Adolphs, H. Bedenbender, D. Dirzus, et. al.: Status Report. Reference Architecture Model Industrie 4.0 (RAMI4.0), Düsseldorf: Verein
Deutscher Ingenieure e.V. / ZVEI – German Electrical and Electronic Manufactures‘ Association, 2015.
[8] N. Gronau, A. Ullrich, G. Vladova: Prozessbezogene und visionäre Weiterbildungskonzepte im Kontext Industrie 4.0, in H. Meier pub.:
Lehren und Lernen für die moderne Arbeitswelt, Gito-Verlag, 2015, pp. 125-143
[9] C. Prinz, F. Morlock, S. Freith, N. Kreggenfeld, D. Kreimeier, B. Kuhlenkötter: Learning Factory modules for smart factories in Industrie
4.0, Procedia CIRP 54, 2016, pp.113-118
[10] E. Abele, J. Metternich, M. Tisch, G. Chryssolouris, W. Sihn, H. ElMaraghy, V. Hummel, F. Ranz: Learning Factories for research,
education, and training, Procedia CIRP 32, 2015, pp. 1-6
[11] M.C. Gwee: Problem-based learning: a strategic learning system design for the education of healthcare professionals in the 21st century,
Kaohsiung J Med Sci 25, 2009, pp. 231–239
[12] N. Harmer, Alison Stokes: The benefits and challenges of project-based learning, School of Geography, Earth and Environmental Sciences,
Plymouth, 2014