The popularity of Augmented Reality (AR) applications has strongly been increased with the worldwide
success of the Pokemon Go videogame released by Niantic in 2016. However, AR offers tangible benefits
in many further areas beyond entertainment, such as advertisement, education, navigation, maintenance,
health, and so on. With the growing spread and success of AR applications in these fields, there has also
been a growing necessity for approaches and technologies for assuring the quality of these applications,
such as testing. A few technologies and frameworks have been recently proposed supporting the
implementation and execution of test scripts that can be used to exercise the applications, but there still
is a lack of effective techniques and tools for the automatic generation of executable test cases. In this
paper, we investigate the possibility of using Model Based Testing techniques to generate executable
test scripts from Finite State Machines modeling the behaviour of the GUI of AR applications, similarly
to other GUI based applications. We have applied several model coverage criteria to design test suites
and we have shown the feasibility of this approach by testing two small example applications involving
Unity3D and Vuforia technologies
Russian Call Girls in Karol Bagh Aasnvi ➡️ 8264348440 💋📞 Independent Escort S...
An Approach for Model Based Testing of Augmented Reality Applications.pdf
1. An Approach for Model Based Testing of
Augmented Reality Applications
Porfirio Tramontana, Marco De Luca
Anna Rita Fasolino
University of Naples “Federico II”
3rd International Workshop on Quality and Measurement of
Software Model-Driven Development - QUAMES 2022
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
2. Motivations
◼ Augmented Reality applications mixes :
◼ A layer provided by a camera
◼ With callbacks related to image recognition
◼ Other layers consisting of interactive widgets
➢ Substantially, they are interactive applications
? How their behaviour can be modelled?
? May Model-based testing be suitable on
them?
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
3. Extended Reality Taxonomy
◼ Virtual reality (immersive or non-immersive VR)
◼ where the application simulates a completely different
environment around the user;
◼ Augmented reality (AR)
◼ where the application combines its own digital
environment with the user’s real-world environment
and allows them to interact with each other.
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
4. Unity 3D / Vuforia Ecosystem
◼ Unity 3D is one of the most popular
alternative for the development of AR and
VR applications
◼ They can be deployed on mobile devices as
native applications
◼ Vuforia provides off-the-shelf services for
continuous image recognition, based on
ML algorithms
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
5. Unity 3D application model
◼ Scene: GUI with a designed layout
(and transparency)
◼ GameObjects: interactive objects
◼ Components: associated to scenes or
game objects, may implement
listeners
◼ Event Listeners: callbacks related to
components
◼ Includes Vuforia callbacks activated by
the recognizing of graphic markers
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
6. Research questions
? How the behaviour of an Augmented
Reality application can be modelled?
? May the model be used to realize Model
Based Testing approaches?
QUAMES @ RCIS 2022 Workshops, Barcelona, May 17, 2022
7. Model Based Testing of AR applications
◼ Modeling → Finite State Machines
◼ Test Selection Criteria → All States
Coverage / All Transitions Coverage / All
Prime Paths Coverage
◼ Test Case Implementation and Execution
→ AirTest
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
8. FSM Model of AR Behaviour
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
9. FSM Model of AR Behaviour
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
10. FSM Model of AR Behaviour
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
11. FSM Model of AR Behaviour
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
12. FSM Model of AR Behaviour
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
13. FSM Model of AR Behaviour
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
14. FSM Model of AR Behaviour
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
15. Test Selection Criteria
◼ All States Coverage
◼ each state of the FSM model is reached by at
least a test case
◼ All Transitions Coverage
◼ each transition of the FSM model is reached
by at least a test case
◼ All Prime Paths Coverage
◼ each prime path on the FSM is covered by at
least a test case.
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
16. AR Test Framework
◼ AirTest
◼ Supports the writing of automatically
executable test scripts
◼ The poco object acts as a starting point to query
the object model of the application during its
execution
◼ Can be used in isolation, without need of a
client/server infrastructure (as AltUnity)
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
17. Test Case Structure
◼ A global object poco
◼ Setup function starting the app and instantiating
poco
◼ A verify function for each state of the model to
assert the reaching of the state by evaluating
widgets reachable via the poco object
◼ A main function with a sequence of state verify
and user event triggering
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
20. Case Studies
◼ Two small Case
Studies:
◼ A1: PointAR
◼ A2: SafariAnimals
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
◼ Three test suites:
◼ TS1: All States Cov.
◼ TS2: All Transitions Cov.
◼ TS3: All Prime Paths Cov.
21. Results
◼ Feasibility of the approach
◼ Transition & Prime Path coverage
produced high code coverage, too
◼ Uncovered branches:
◼ An unfeasible branch
◼ A branch that needed a larger path
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
22. Perspective, Limitations, Future Work
◼ Preconditions
◼ AirTest currently does not support setting of
image markers
◼ To have it an interaction with an external image
injector should be considered
◼ Model formalization
◼ Need to generalize the application model
◼ to reverse engineer it from code
◼ To generate test cases from the model
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
23. Perspective, Limitations, Future Work
◼ Need to evaluate on larger applications
and with real defects
◼ There are very few examples of open source
AR projects with test cases and documented
faults
◼ Mutants could be considered
QUAMES @ RCIS 2022, Barcelona, May 17, 2022
24. Thank you for your attention!
QUAMES @ RCIS 2022, Barcelona, May 17, 2022