TY - JOUR
T1 - Safety in Autonomous Driving Decision-Making
T2 - A Review of Challenges, Methods, and Technological Evolution
AU - Xu, Jiahui
AU - Shu, Yiming
AU - Tang, Jiwei
AU - Zhong, Jiaru
AU - Wang, Hong
AU - Wang, Weida
AU - Sun, Chen
N1 - Publisher Copyright:
© 2000-2011 IEEE.
PY - 2026
Y1 - 2026
N2 - Although autonomous driving technology holds great promise for eliminating human errors, safety issues remain a significant challenge and may hinder the future development of intelligent transportation systems. In the decision-making domain, the progress of safety assurance has lagged behind the rapid advancements in autonomous driving technologies in recent years. Therefore, this review aims to examine safety assurance methods within decision-making systems. This paper provides a comprehensive review of safety in autonomous driving decision-making by categorizing major safety challenges, summarizing common mitigation strategies, and discussing their evolution alongside the development of decision-making technologies in autonomous vehicles. Then, the scope is extended to the vehicle-to-everything perspective, presenting safety-oriented decision-making methods in connected environments. Then, based on the above classification of safety algorithms, this paper develops a plug-in safety decision tool and constructs a benchmark to demonstrate the safety gains achieved by different algorithms. Finally, the paper identifies existing research gaps and outlines potential directions for future work. The review seeks to raise awareness of safety issues in decision-making and serve as a reference for researchers interested in advancing safety in this field. The benchmark implementation is available at https://github.com/salt0107fish/DMS-benchmark
AB - Although autonomous driving technology holds great promise for eliminating human errors, safety issues remain a significant challenge and may hinder the future development of intelligent transportation systems. In the decision-making domain, the progress of safety assurance has lagged behind the rapid advancements in autonomous driving technologies in recent years. Therefore, this review aims to examine safety assurance methods within decision-making systems. This paper provides a comprehensive review of safety in autonomous driving decision-making by categorizing major safety challenges, summarizing common mitigation strategies, and discussing their evolution alongside the development of decision-making technologies in autonomous vehicles. Then, the scope is extended to the vehicle-to-everything perspective, presenting safety-oriented decision-making methods in connected environments. Then, based on the above classification of safety algorithms, this paper develops a plug-in safety decision tool and constructs a benchmark to demonstrate the safety gains achieved by different algorithms. Finally, the paper identifies existing research gaps and outlines potential directions for future work. The review seeks to raise awareness of safety issues in decision-making and serve as a reference for researchers interested in advancing safety in this field. The benchmark implementation is available at https://github.com/salt0107fish/DMS-benchmark
KW - Autonomous driving
KW - artificial intelligence
KW - decision-making
KW - safety
UR - https://www.scopus.com/pages/publications/105053689691
U2 - 10.1109/TITS.2026.3740988
DO - 10.1109/TITS.2026.3740988
M3 - Article
AN - SCOPUS:105053689691
SN - 1524-9050
JO - IEEE Transactions on Intelligent Transportation Systems
JF - IEEE Transactions on Intelligent Transportation Systems
ER -