Detection and diagnosis of deviations in distributed systems of autonomous agents

Abstract

Given the complexity of cyber-physical systems (CPS), such as swarms of drones, often deviations, from a planned mission or protocol, occur which may in some cases lead to harm and losses. To increase the robustness of such systems, it is necessary to detect when deviations happen and diagnose the cause(s) for a deviation. We build on our previous work on soft agents, a formal framework based on using rewriting logic for specifying and reasoning about distributed CPS, to develop methods for diagnosis of CPS at design time. We accomplish this by (1) extending the soft agents framework with Fault Models; (2) proposing a protocol specification language and the definition of protocol deviations; and (3) development of workflows/algorithms for detection and diagnosis of protocol deviations. Our approach is partially inspired by existing work using counterfactual reasoning for fault ascription. We demonstrate our machinery with a collection of experiments.

ACAV: A Framework for Automatic Causality Analysis in Autonomous Vehicle Accident Recordings

“…Biebl et al [14] presented a causal model to predict accident risks in an intersection for drivers with impairments. In addition to works focusing on AI [11,18,19,59], some works have applied causality to the security analysis of CPSs [34,36,37,41,42]. Zhang et al [60] monitored, inspected, and located anomalies in industrial control systems using a causal model based on maximum information coefficient and transfer entropy.…”