A survey on verification strategies for intelligent transportation systems

摘要

As intelligent systems are increasingly entering everyday life, in domains such as transportation, resource distribution, health care, or retail, developing suitable verification mechanisms for such systems becomes vital. From a formal point of view, the employed intelligent sensor actuator systems (ISAS) constituting such intelligent systems combine three different technologies: control systems, distributed systems, and learning and reasoning. While each of the parent domains features tested and proven verification methods, simply combining the tasks unfortunately leads to a combinatorial explosion of complexity. This paper presents an overview and classification of currently employed techniques for handling ISAS in terms of: cyber-physical systems, intelligent autonomous robots, or intelligent agents. The article argues that each of the three classical perspectives misses one important characteristic of ISAS and proposes to combine the three for a full solution. The paper argues that in particular two mechanisms are promising: an intelligent environments perspective that verifies local safety and techniques for context-aware monitoring that allow a mobile system to leverage context-awareness to reduce complexity for self-monitoring tasks.