=Paper=
{{Paper
|id=Vol-2815/CERC2020_paper13
|storemode=property
|title=Integration of Pattern-Based Learning Management Systems With Immersive Learning Content
|pdfUrl=https://ceur-ws.org/Vol-2815/CERC2020_paper13.pdf
|volume=Vol-2815
|authors=Michael Winterhagen,Dominic Heutelbeck,Benjamin Wallenborn,Matthias L. Hemmje
|dblpUrl=https://dblp.org/rec/conf/cerc/WinterhagenHWH20
}}
==Integration of Pattern-Based Learning Management Systems With Immersive Learning Content==
https://ceur-ws.org/Vol-2815/CERC2020_paper13.pdf
Visual Computing and Data Analytics
Integration of Pattern-Based Learning
Management Systems with Immersive Learning
Content
Michael Winterhagen1[0000−0002−9487−3544] , Dominic
Heutelbeck2[0000−00003−1395−2775] , Benjamin Wallenborn3[0000−0001−9234−5310] ,
and Matthias L. Hemmje1[0000−0001−8293−2802]
1
Chair of Multimedia and Internet Applications (MMIA), Faculty of Mathematics
and Computer Science, University of Hagen, Germany {michael.winterhagen,
matthias.hemmje}@fernuni-hagen.de
2
Research Institute for Telecommunication and Cooperation, Germany
dheutelbeck@ftk.de
3
University of Hagen, Center for Media and IT (ZMI), Germany
benjamin.wallenborn@fernuni-hagen.de
Abstract. In our previous e-learning publications we introduced the
concepts of Course Authoring Tools (CAT), and of Didactical Structural
Templates (DST) which are a further development of the CAT. DSTs are
defined as a possibility to describe the pedagogical structure of a course,
a study program or an applied game in an abstract way. The idea of DSTs
is based on the structure of IMS Learning Design (IMS-LD), which is a
quasi-standard for modelling learning structures. We have shown what
DSTs are useful for and that there is a need for an editor for DSTs which
we already presented in combination with the Didactical Structural Tem-
plate Manager (DSTM). In this paper we will focus on and go in deeper
detail in implementing the same DST as a classic Learning Management
(LMS) course, as a LMS course with applied gaming content and as a
stand alone applied game. Therefore, these scenarios will show that it
will be possible for learners to switch between different implementations
of a specific DST and the learners having the same learning progress as if
they had used just one of the implementations of this specific DST. One
main contribution is to demonstrate the integration between, web-based,
immersive Virtual Reality learning activities developed as native appli-
cations, and VR authoring tools by a gateway server. In addition, we
present an RESTful API for sharing the DST and the used Competence
Profiles to external tools.
Keywords: IMS Learning Design ➲ IMS-LD ➲ Didactical Structural Tem-
plates ➲ Learning Tools Interoperability ➲ LTI.
1 Introduction
Producing learning content for Higher Education Institutes (HEI) consumers
is quite complex in Learning Management Systems (LMSs) for producers of
Copyright © 2020 for this paper by its authors.
Use permitted under Creative Commons License 211 CERC 2020
Attribution 4.0 International (CC BY 4.0).
Visual Computing and Data Analytics
2 M. Winterhagen et al.
learning content. The Knowledge-Management Ecosystem Portal (KM-EP) [6]
is an Educational Portal which contains different management systems. A Course
Authoring Tool (CAT) has been introduced in [16] within the KM-EP to reduce
this complexity and make it simple to produce learning content in the way that
the producers only have to fill in the necessary information and are able to
concentrate in producing the learning content instead of configuring the learning
content within the LMS.
1.1 Motivation, Problem Statement and Approach
The so-called Didactical Structural Templates (DST) have been introduced in
[19] and extended in [17]. As described, the DSTs are based on the IMS Learning
Design (IMS-LD) [8] and represent the pedagogical structure of a course and
cannot only be used as pedagogical structure for creating courses. In fact, the
DSTs can also be used as a pedagogical structure for a hybrid environment
existing of a ”classical” course with integrated applied gaming content just like a
pedagogical structure for applied game which can be a web-based computer game
or a Virtual Reality (VR)/ Augmented Reality (AR) based game. Therefore, one
DST can have different implementations.
The advantage of this approach is, that learners will be able to switch between
different implementations of one DST whenever they want to and they have got
the same learning progress as if they had used only one specific implementation
of this DST. This means if learners like gaming, they can use the applied gaming
implementation to work on the learning content. If it is easier for the learners to
answer the self-tests or the final test – to stay in the exemplary stated pedagogical
structure of a course – as e.g. multiple-choice quizzes, they can switch to a course
within an LMS to answer the questions.
Within a wide rage of domains, realizing learning activities as immersive
experiences, e.g., in VR, may significantly improve the learners experience and
success [1]. While tools like the WebXR Device API [15] have the potential to
close the gap between traditional browser-based learning and content played out
via dedicated VR Hardware. In industrial VR learning environments, specialized
hardware-based controllers may be necessary for the appropriate simulation of
machine operation, and the virtual environments are of high fidelity, derived
from engineering CAD systems. Complex environments, such as models of com-
plete factories are very demanding with regards to graphics capabilities of the
hardware. Thus, learning activities and self tests in VR often require the use
of native application running on a capable machine, connected to specialized
controllers. While CATs are often also able to support the authoring of web-
based content, immersive applications, both native and WebXR-based, require
specialized authoring tools which are able to handle the spatial and graphical
content and complex interfaces. In practice, such authoring tools are built on
top of game development tools and platforms, such as the Unity [14] or Unreal
[5] engine.
Such immersive activities are potentially embedded within a DST alongside
traditional web-based content. The integration of this type of content into the
CERC 2020 212
Visual Computing and Data Analytics
Integration of Pattern-Based Learning Management Systems 3
authoring process, deployment and execution of a course instance implementing
a DST rises a number of technical and architectural challenges.
During web-based learning in a traditional LMS, the use of external web-
based tools is well supported by the Learning Tools Interoperability (LTI) [10]
Standard by a number of modular APIs and protocols. This allows to start activ-
ities implemented using WebGL or WebXR directly. However the reliance of LTI
on OAuth [13], Transport Layer Security (TLS) [7], and forwarding between tool
server and LMS server for launch requests does prohibit directly launching into
a native application installed on the users computer, while retaining authentica-
tion and course context information. E.g., launch URLs have to be public TLS
secured web-resources. A locally running tool cannot fulfill the TLS requirements
running on ’localhost’.
To summarize the remainder of this paper addressing our five research ques-
tions, which we will require to work on shown below:
1. How can a DST representation be applied to a “classical” course production
in an LMS?
2. How is the responsibility of content creation distributed between the different
content authoring tools.
3. How can we share and cross-reference content between the different tools
and execution environments.
4. Can LTI-style launch requests be realized in order to launch into native
applications while retaining the learners authentication and course context?
5. How can native external tools send back results to the LMS?
1.2 Methodology
As the basis of our research methodology, the multi-methodological framework
of Nunamaker and Chen [12] is used for the structured research and development
of information systems. The framework is divided into four phases supporting
different methodological strategies: Observation, Theory Building, System De-
velopment, and Experimentation. To achieve our research goal to answer our
research questions, the methodological phases can be executed repeatedly in
any order. It is also possible to return to previous phases.
2 Concepts and Technologies
LTI is introduced in [11]. LTI supports a connection between LMSs and external
applications like gaming platforms. An applied game prototype based on the
Unity Game Engine (Unity) [14] has been introduced in [18]. This prototype
is later used as an example for implementing an LTI connection. However, this
only works for web-applications and not for desktop-applications. To enable this
also for desktop applications, the solution presented in [18] has to be extended.
213 CERC 2020
Visual Computing and Data Analytics
4 M. Winterhagen et al.
3 Conceptual Work
After describing the problems and research questions in chapter 1.1 we want to
present our conceptual work in this chapter. To realize the DST as an applied
game we have different challenges we have to work on, to make this option work:
1. LTI can only be used for web-based applications, not for stand alone or
desktop-applications. How can LTI be used for our scenario?
2. How can the applied game access the DST?
3.1 LTI Launch Requests and Services for Immersive Native
Applications
In this section we present our architecture (see figure 1) and implementation
of a technical solution for the above described challenges arising from the need
for Immersive Native Applications (INAs) in order to support immersive VR
learning content. Overall, the idea is to use a so-called LTI Gateway to mediate
between a) the LMS implementing LTI, b) the VR-specific authoring tools, and
c) the VR learning tool installed on the users computer. The goal is to support
standard LTI 1.3 [10] launch requests to trigger learning activities using the local
INA and to allow it to use LTI Advantage [11] services offered by the platform
(LMS), e.g. Assignment and Grade Services 2.0 [9].
In LTI 1.3 a platform and an external tool have to be carefully paired and
the tool and platforms have to be authorized by the respective administrators.
In this process, public keys and several URLs have to be exchanged and unique
IDs for tool and platform have to be provisioned. Generally the URLs have to be
mutually reachable by the systems hosting either component and it is mandatory
for both sides to have all connections secured by TLS supported by the matching
certificates.
In order to use the INA, it has to be installed locally on the machine used
by the learner. While in special circumstances, this may be a dedicated ma-
chine where it is possible to assign a DNS entry and to obtain a TLS certificate.
Then, the INA could directly support the LTI protocols by running matching
background services. However, in practice learners will not use a singe machine
shared among them. Thus, even in a learning environment with multiple ma-
chines that could fulfill these conditions, in the LTI domain model each machine
would constitute a different external tool which implies that in a course the
same learning activity on different machines would be listed separately. Further
more, it is likely that a course is not consumed by a predetermined number of
local learners, but by a variable number of remote learners with machines in
uncontrolled environments and behind NAT routers. Thus, a direct provisioning
of local installations of the INA as LTI tools is not feasible.
In order to use other LTI Advantage services, external tools acquire OAuth
2.0 [13] access tokens. While the OAuth 2.0 standard technically allows for INAs
to obtain such tokens directly, it is not guaranteed, that the matching OAuth
CERC 2020 214
Visual Computing and Data Analytics
Integration of Pattern-Based Learning Management Systems 5
flows are supported by all LMS. However the flows for web-based applications
to obtain the tokens is mandatory for the LTI standards.
To support both launch request and and LTI Advantage services, a dedicated
web-based service, the LTI Gateway is introduced. The LTI Gateway itself acts
as a single point of entry and an LTI tool for the platform and is able to relay
the launch request to an INA and service calls from the INA to the platform.
Thus, pairing between tool and platform only happens once for the gateway and
not for each local installation of the INA.
The gateway can be referenced in a course like any other external LTI tool.
When a user decides to access the learning activity and clicks on the matching
link provided by the platform, a standard LTI Launch Request to the gateway
is triggered. In the process of the launch request protocol, the gateway obtains
information about the user and the context of the launch via different claims
stored in the JSON Web Token (JWT) [4] used in the launch. In addition, the
gateway obtains an access token to use the platforms LTI Advantage services.
The course author may have decided to add additional custom parameters (string
key-value pairs) to the launch, which are also made available to the tool in the
matching JWT claim.
Given, that users may have different sections and activities of a course open
at the same time or that users may, often unintentionally, click a launch link
multiple times the gateway has both to disambiguate launch requests based on
the user and context information.
Thus, the gateway is a stateful service which upon receiving a launch request
and obtaining the information creates a gateway session identified by the tuple
of user identifier and LTI context. Each session has a predefined, decided by the
gateway administrator, time to live (TTL) and a randomly generated Base64
encoded 128-bit Universally Unique Identifier (UUID), acting as a session key.
For this session, the triple, session key, LTI Advantage access token, and JWT
token are persistent in RAM for the TTL. Before creating a new session for a
launch request, it is checked if another session for the given triple already exists,
which is used if present.
After creating, or loading the session, the learner is presented with a TLS
secured web page containing information on the origin of the launch and instruc-
tions on how to obtain the INA, if the user still needs to install it. The user is
presented with a launch link to start the INA. The launch link is structured as
follows: [CUSTOM PROTOCOL]:[TOOL NAME]?sessionKey=[SESSION KEY], e.g.,
i2l:virtualFactory?sessionKey=jhSLqIopUf ZiP86n6VDaA. At install time,
the INA registers itself as the handler for the custom protocol (e.g., adding the
respective registry entries for Windows). Thus, when clicking the link, the web
browser will launch the INA and provide the URL as a command line argument.
In this way, the INA is able to obtain the session key. Compared to a direct
authorization of the INA and its user via OAuth this opens an attack vector,
where on a compromised machine the launch may be intercepted and the session
key may be obtained by a malicious application. This however is only possible
as a local attack and the key is always encrypted in transit over the network.
215 CERC 2020
Visual Computing and Data Analytics
6 M. Winterhagen et al.
We consider this to be an acceptable remaining risk. This requires local code
execution privileges which imply further extensive attack potential.
Given the session key, the INA can retrieve the session data from the gateway
by accessing GET: /launchRequest/{sessionKey}, returning a JSON
object containing user information, LTI resource link, line item, and context
and custom data provided by the course author.
The INA can now start the learning activity according to this information.
For each LTI Advance service the gateway also provides matching proxy services
which accept the session key as authentication, such as results or scoring services.
Whenever the INA accesses the proxy services with the session key, internally
the request is mapped to a matching LTI request to the platform, using the
stored access token, which is never shared with the INA.
3.2 VR Authoring and Resource Server
Figure 1 gives a rough overview of the interaction of the different parts described
in the previous sections.
Fig. 1. LTI Gateway and VR Authoring Tools
Within the VR Authoring Tool, the producers can define the Training Se-
quences, Compendium Entries and Media Assets which are published on a server
which holds this content. The meaning of this elements are as follows:
The Media Assets can be texts or pictures and can be used by Compendium
Entries.
The Compendium Entries can be understood as elements of knowledge trans-
fer. They consist of contents in a specific order. Each content is composed of e.g.
CERC 2020 216
Visual Computing and Data Analytics
Integration of Pattern-Based Learning Management Systems 7
texts and pictures which are stored as Media Assets. The Compendium Entries
e.g. deliver information about the environment of the game or the Training Se-
quences. To access the Media Assets, the Compendium Entry has to connect to
the content server and download it from there.
At last, the Training Sequences can be understood as learning elements of a
learning path. This Training Sequences consist of so-called Tasks which are in a
specific order and have to be done in the defined order. Each Task contains so-
called Objectives which have no specific order. Each Objective will be displayed
in a menu in the game. To complete a task, each Objective has to be done
completely.
Having described the Training Sequences and its structure, we can define a
mapping between the IMS-LD and an applied game using the Training Sequences
as shown in table 1.
Table 1. Mapping IMS-LD - applied game
IMS-LD element applied game element
method applied game
play levels
act Training Sequence
activity structure Task
activity Objective
3.3 Delivering the Structure of a Didactical Structural Template to
an External Application
In this section we present our technical solution for the above described challenge
2. We support two kinds of delivering the structure of an DST to an external
application:
1. The Didactical Structural Template Manager (DSTM) provides the possi-
bility to export an DST as XML-file which meets the IMS-LD specification.
Therefore, the DST could be exported in this format and can be used to
design the structure of the game.
2. The KM-EP provides a Learning Design API which will be described in short
in the following. The external application can access this Learning Design
API to get all information about the needed DST.
The Learning Design API is defined as a Representational State Transfer (REST)
[3] with the following endpoints:
– GET: /webservice/learningdesign
This endpoint returns a list of all available DSTs.
– GET: /webservice/learningdesign/{id}
This endpoint returns the structure of a specific DST.
217 CERC 2020
Visual Computing and Data Analytics
8 M. Winterhagen et al.
The return value of each endpoint is made in JavaScript Object Notation (JSON)
[2]. The specific return value’s format will not be explained in detail here. With
this both kinds of delivering the DST to an external application, it is possible
for e.g. an applied game to access the needed DST.
4 Prototypical Implementation
After describing the technical requirements and the technical prerequisites, we
are now ready to show, how it is possible to implement a DST as
1. a normal Moodle course,
2. a gamified Moodle course, and
3. an applied game.
This will be done within the following subsections.
4.1 Implementing a Didactical Structural Template as a Moodle
Course
After defining an DST at first we want to implement this into a Moodle course.
To do so, we have to go to the DSTM within the KM-EP and press on the install
button next to the DST we want to implement.
After pressing the install button, the DSTM will implement the selected DST
as Moodle course according to the mapping we presented in [17] with default
values for each learning element. Therefore, we now have an empty Moodle course
created which we have to fill with the needed learning content.
To fill this empty Moodle course with learning content, we can now use the
CAT within the KM-EP to edit every learning element of the created Moodle
course.
After saving the changed learning element, the changes will directly be trans-
ferred to Moodle. After the complete course has been filled with the needed
learning content, the course is ready to be published and used by learners.
4.2 Implementing a Didactical Structural Template as a Gamified
Moodle Course
As second option we want to implement the same DST of 4.1 as a Moodle course
but with an applied game.
Before we implement the course, there has to be implemented an applied
game with an LTI interface. For our example, we use a web-based applied game
implemented with the Unity Engine (Unity) [14].
We already described in detail in [18] how the LTI connection between Moo-
dle and the applied game works. We will show exemplary, how the implemented
applied game can be included within an Moodle course.
At first the URL of the created applied game has to be added as software
into the KM-EP.After the applied game is known within the KM-EP, it can be
CERC 2020 218
Visual Computing and Data Analytics
Integration of Pattern-Based Learning Management Systems 9
referenced by a Moodle course. Therefore, an learning element has to be defined
as asset which references the applied game with help of the CAT.
When a learning element with an underlying applied game is selected by the
learners, the applied game will start and the learning result will be transferred
back to the LMS via LTI.
4.3 Implementing a Didactical Structural Template as an Applied
Game
As third option we want to implement the same DST of 4.1 as a stand alone
applied game. In our scenario, this game will be an VR/AR game.
Prototypical Implementation of LTI connections for Stand Alone Ap-
plications To make it possible to call the LTI Gateway mentioned in 3.1, we
have to define an LTI connection in Moodle (see figure 2).
Fig. 2. Define LTI connection in Moodle.
For our purpose we need at least version 1.3 of the LTI specification. After
the LTI connection for the LTI Gateway has been defined within Moodle for a
specific learning element, a click on the corresponding learning element will open
a website which provides a link to the applied game.
After clicking on the provided link, an applied game will be downloaded and
started. This applied game will contain the LTI connection to return learning
results to Moodle.
5 Conclusions
In this paper we have shown how to implement a given DST
– as a normal Moodle course,
– as a gamified Moodle course, and
– as an applied (VR) game without Moodle or another LMS.
219 CERC 2020
Visual Computing and Data Analytics
10 M. Winterhagen et al.
We also have extended the LTI connection, which until now only could be used
for web-based applications, for stand alone applications via an LTI Gateway. We
used this LTI Gateway for an applied game which implemented an DST.
To give an external application access to the pedagogical structure of an DST
we offered two options:
1. exporting the pedagogical structure of an DST via the extended IMD-LD
specification and
2. providing a REST-API which provides the pedagogical structure of an DST
in JSON format.
If we have two different DSTs, namely DST A and DST B, the learners
learning progress can be further developed, when the learners switch between
different implementations of DST A. If they switch into an implementation on
DST B, they will have anoter, separate learning progress apart from DST A.
5.1 Future Work
Future work will be a further formal evaluation of the DSTs and the DSTM,
especially in the way to use the DST as pedagogical structure for an applied
game without connecting it to a Moodle course. Also a formal further evaluation
will be based on learners switching between different implementations of an DST
having the same learning progress.
The process of creating a Moodle course by installing it from an DST is not
as good as it could be. To improve the installation process, it would be desirable
to have a wizard, so that the learning content can already be filled in while
installing the DST as a course.
We also have to extend the possibility to connect to applied games via LTI
for desktop-applications.
Acknowledgements
This publication has been produced in the context of the RAGE and Immerse2Learn
projects. The projects have received funding from the European Union’s Hori-
zon 2020 Research and Innovation Action under grant agreement No 644187 and
from European funds for regional development (EFRE). However, this paper re-
flects only the authors’ views and the European Commission is not responsible
for any use that may be made of the information it contains.
References
1. Allcoat, D., von Mühlen, A.: Learning in virtual reality: Effects on per-
formance, emotion and engagement. Research in Learning Technology 26
(2018). https://doi.org/https://doi.org/10.25304/rlt.v26.2140, https://journal.
alt.ac.uk/index.php/rlt/article/view/2140
CERC 2020 220
Visual Computing and Data Analytics
Integration of Pattern-Based Learning Management Systems 11
2. Crockford, D.: The javascript object notation (json) data interchange format
(2017), https://tools.ietf.org/pdf/rfc8259.pdf
3. Fielding, R.T.: Architectural Styles and the Design of Network-based Software
Architectures. Ph.D. thesis (2000), https://www.ics.uci.edu/~fielding/pubs/
dissertation/fielding_dissertation_2up.pdf
4. Force, I.E.T.: Rfc 7519 - json web token (jwt) (2015), https://tools.ietf.org/
html/rfc7519
5. Games, E.: The most powerful real-time 3d creation platform - unreal engine
(2020), https://www.unrealengine.com/en-US/
6. Globit-GmbH: Ecosystem portal, https://www.globit.com/#ep
7. IETF: The tls protocol (1999), https://tools.ietf.org/html/rfc2246
8. I.G.L.-Consortium: Learning design specification (2003), https://www.imsglobal.
org/learningdesign/index.html
9. I.G.L.-Consortium: Ims learning tools interoperability (lti) assignment and grade
services specification (2019), https://www.imsglobal.org/spec/lti-ags/v2p0/
10. I.G.L.-Consortium: Learning tools interoperability core specification (2019), http:
//www.imsglobal.org/spec/lti/v1p3/
11. I.G.L.-Consortium: Lti v1.3 and lti advantage (2019), https://www.imsglobal.
org/activity/learning-tools-interoperability
12. Nunamaker, J.F., Chen, M., Purdin, T.M.D.: System development in infor-
mation systems researchs (1990), http://gkmc.utah.edu/7910F/papers/JMIS%
20systems%20development%20in%20IS%20research.pdf
13. OAuth: Oauth 2.0 (2012), https://oauth.net/2/
14. Unity: Unity technologies, https://unity3d.com/de
15. W3C: Webxr device api (2019), https://www.w3.org/TR/webxr/
16. Wallenborn, B.: Entwicklung einer innovativen Autorenumgebung für die uni-
versitäre Fernlehre. Ph.D. thesis, FernUniversität Hagen (2018), https://ub-
deposit.fernuni-hagen.de/receive/mir_mods_00001428
17. Winterhagen, M., Hoang, M.D., Lersch, H., Fischman, F., Then, M., Wallenborn,
B., Heutelbeck, D., Fuchs, M., Hemmje, M.: Supporting structural templates for
multiple learning systems with standardized qualifications. In: EDULEARN20 Pro-
ceedings (2020)
18. Winterhagen, M., Salman, M., Then, M., Wallenborn, B., Neuber, T., Heutelbeck,
D., Fuchs, M., Hemmje, M.: Lti-connections between learning management systems
and gaming platforms. Journal of Information Technology Research 13 (2020)
19. Winterhagen, M., Then, M., Wallenbrn, B., Hemmje, M.: Towards structural
templates for learning management systems taking into account standardized qual-
ifications. Tech. rep., FernUniversität Hagen (2019), https://www.researchgate.
net/publication/340601087_Towards_Structural_Templates_for_Learning_
Management_Systems_taking_into_account_Standardized_Qualifications
221 CERC 2020