The paper addresses the problem of online monitoring of real-time systems, where the communication between the monitored system and the monitor is subject to unknown delays. The authors propose a symbolic, zone-based monitoring algorithm that can handle such delays, which consist of a constant latency and a bounded jitter.
The key highlights and insights are:
Online monitoring can be achieved by compiling the specification (expressed in Metric Interval Temporal Logic) into a timed Büchi automaton and then symbolically executing the automaton on the observed trace of the system. However, this approach is correct only if the actions of the monitored system can be observed immediately by the monitor.
In practice, there is usually a communication delay between the system and the monitor, induced by various types of circuitry at their interfaces. The authors present a symbolic monitoring algorithm that can handle such parametric delays without resorting to costly verification procedures for parameterized timed automata.
The algorithm uses a zone-based representation to compute the set of states that are reachable by possible "ground-truth" timed words that are consistent with the observed trace, given the bounds on the latency and jitter.
The algorithm can provide exact verdicts (satisfaction, violation, or inconclusive) under the unknown delay, and can also report the set of delays that are consistent with the observed trace and the specification (or its negation).
The authors provide a prototype implementation of the algorithm, called MoniTAal, and evaluate it on a case study involving a gear controller model.
To Another Language
from source content
arxiv.org
Deeper Inquiries