Back to Search
Start Over
Anomaly diagnosis of connected autonomous vehicles: A survey.
- Source :
-
Information Fusion . May2024, Vol. 105, pN.PAG-N.PAG. 1p. - Publication Year :
- 2024
-
Abstract
- • Survey of CAV anomaly diagnosis from anomaly detection and anomaly interpretation. • Taxonomies of anomaly detection based on concept of normality and CAV applications. • Clarification of the concept of anomaly interpretation. • Anomaly interpretation combining the analysis of road vehicle safety risks. • Comprehensive and detailed review of recent advances for CAV anomaly diagnosis. Connected autonomous vehicles (CAVs) are revolutionizing the development of transportation due to their potential to improve transportation performance in many ways, such as enhanced traffic mobility, road compacity, operation safety, and environmental sustainability. Nevertheless, the issue of road vehicle safety in CAVs remains to be fully solved. Data collected from multiple sources provide information about the internal status of the system and the situation of its surroundings, and the occurrence of data anomalies indicates the existence of potential safety risks. Thus, anomaly diagnosis is of major importance to analyze the nature or cause of underlying safety risks and provide insightful information for the subsequent decision-making that ensures safety. Anomaly diagnosis consists of two basic tasks: anomaly detection and anomaly interpretation. In this paper, both of these two tasks are comprehensively surveyed. For anomaly detection, the following four aspects are covered: 1) formalized definition of an anomaly, 2) classification of anomalies, 3) taxonomies of anomaly detection techniques, and 4) review of recent advances for anomaly detection in CAV applications. For anomaly interpretation, related works are investigated in the context of 1) the anomaly detection process, and 2) the tested/monitored system/process, respectively. The novelty particularly lies in the latter, where the interpretation of anomalies combining the analysis of road vehicle safety risks is presented, and related works for anomaly interpretation in CAV applications are reviewed by analyzing 1) functional safety risks, 2) safety of the intended functionality (SOTIF) risks, and 3) cyber security risks, respectively. Finally, open issues, challenges, future directions, and emerging technologies for anomaly diagnosis in CAVs are discussed. [ABSTRACT FROM AUTHOR]
Details
- Language :
- English
- ISSN :
- 15662535
- Volume :
- 105
- Database :
- Academic Search Index
- Journal :
- Information Fusion
- Publication Type :
- Academic Journal
- Accession number :
- 175243596
- Full Text :
- https://doi.org/10.1016/j.inffus.2024.102223