Reliability and Safety of Autonomous Systems Based on Semantic Modelling for Self-Certification

A novel modelling paradigm for online diagnostics and prognostics autonomous systems is presented. model the system being diagnosed designed using a logic-based formalism. The supports run-time ability to verify that safe reliable operation within dynamic environment. based on philosophy there are different types of semantic relationships between states parts system. finite state automaton devised each sensed component some non-sensed components. To capture interdependencies components such complex robotic platform, automatons were related other by relationships. Modality was utilised formalism abstract add measures possibility uncertainty complexity analysed evaluate its scalability applicability systems. results demonstrate not linear computational time 10 ms required achieve 2200 KB knowledge interdependences. detect mitigate hardware failures demonstrated confined space operation. Our findings provide evidence our approach significant challenge safety compliance reliability in