Page 40 - Read Online

Ernest et al. Complex Eng Syst 2023;3:4 I http://dx.doi.org/10.20517/ces.2022.54 Page 21 of 22

proven to be adherent to these safety specifications. The resulting GFT is then guaranteed to be adherent to
specifications over all input values while being explainable and performant. While this study does not intend
to demonstrate performance of an entire Starcraft 2 game controller, it demonstrates the capability of a Fuzzy
Logic-based AI system to be trained and proven to adhere to safety specifications in a specific subset of the
control actions within this game that represent mission/safety-critical elements.

DECLARATIONS
Authors’ contributions
Made contributions to conception and design of the work, developed most associated code for Starcraft inte-
gration and reinforcement learning, performed data analysis and figures of interest, manuscript writing and
related tasks: Ernest N
Contributed to parts of the conception and design, implemented the formal verification techniques towards
Fuzzy Systems, performed analysis and translation of formal specifications, manuscript writing and related
tasks: Arnett T
Created the interfaces needed to easily integrate the Thales’s toolkit with Python environments such as the SC2
interfaces used in this work, manuscript writing and related tasks: Phillips Z

Availability of data and materials
While the resultant GFT AI model and the Psion and EVE software cannot be shared openly, the Starcraft 2
map files utilized in the scenarios shown in this study could be. They are not currently hosted, but can be made
available upon request.

Financial support and sponsorship
None.

Conflicts of interest
All authors are employees of Thales Avionics Inc., from which two software packages that are commercially
available were utilized in this research.

Ethical approval and consent to participate
Not applicable.

Consent for publication
Not applicable.

Copyright
© The Author(s) 2023.

REFERENCES
1. Zhao Y, Wang H, Xu N, Zong G, Zhao X. Reinforcement learning-based decentralized fault tolerant control for constrained interconnected
nonlinear systems. Chaos, Solitons & Fractals 2023;167:113034. DOI
2. Tang F, Niu B, Zong G, Zhao X, Xu N. Periodic event-triggered adaptive tracking control design for nonlinear discrete-time systems via
reinforcement learning. Neural Netw 2022;154:43-55. DOI
3. Silver D, Huang A, Maddison CJ, et al. Mastering the game of Go with deep neural networks and tree search. Nature 2016;529:484-9.
DOI
4. Gunning D, Aha D. DARPA’s explainable artificial intelligence (XAI) program. AIMag 2019;40:44-58. DOI
5. Ross TJ. Fuzzy logic with engineering applications. John Wiley & Sons; 2009. DOI
6. Castro JL. Fuzzy logic controllers are universal approximators. IEEE Trans Syst, Man, Cybern 1995;25:629–35. DOI
7. Buckley J, Siler W, Tucker D. A fuzzy expert system. Fuzzy Sets and Systems 1986;20:1–16. DOI