Enhancing reliability analysis of AI systems using STPA: a case study on the DroneResponse UAV system

Cong Pan; Yan Gao; Ai Gu

doi:10.1117/12.3011516

7 December 2023 Enhancing reliability analysis of AI systems using STPA: a case study on the DroneResponse UAV system

Cong Pan, Yan Gao, Ai Gu

Proceedings Volume 12941, International Conference on Algorithms, High Performance Computing, and Artificial Intelligence (AHPCAI 2023); 129415E (2023) https://doi.org/10.1117/12.3011516
Event: Third International Conference on Algorithms, High Performance Computing, and Artificial Intelligence (AHPCAI 203), 2023, Yinchuan, China

Abstract

The Systems-Theoretic Process Analysis (STPA) method has been widely applied in safety-critical engineering scenarios. Recently, scholars have made preliminary explorations of using STPA for reliability analysis and have made initial progress. However, there lacks a comprehensive, multi-layer perspective of control structure modeling and combined view of reliability of AI systems via STPA We propose an STPA method called STPA-AIR, which is oriented towards a more comprehensive view of AI system reliability. Starting from the critical task paths of the system, we establish the system control structure and analyze the control behaviors and causal scenarios that may lead to critical task failures for reliability analysis. The application of this method on the DroneResponse Unmanned Aerial Vehicle (UAV) system demonstrates the effectiveness of the research.

(2023) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Cong Pan, Yan Gao, and Ai Gu "Enhancing reliability analysis of AI systems using STPA: a case study on the DroneResponse UAV system", Proc. SPIE 12941, International Conference on Algorithms, High Performance Computing, and Artificial Intelligence (AHPCAI 2023), 129415E (7 December 2023); https://doi.org/10.1117/12.3011516

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available