Abstract
Runtime monitoring is a key element in safety critical flight control applications, where fault tolerance is often achieved by automatic failure detection and subsequent reconfiguration. Model-based runtime monitoring uses knowledge of the system dynamics in order to detect incompliant behavior. In this paper, a novel model structure for model-based monitoring of closed-loop systems is proposed, that is inspired by an extended incremental nonlinear dynamic inversion control approach. It ensures that the dynamics of the monitoring models reflect the relative degree of the plant including its actuator dynamics. The model structure thereby allows decoupling of residual signals from the control input as the lag in the response to control inputs is correctly incorporated in the model. Monitoring of allowable intervals for dynamic responses is enabled by integration of the model structure with a method for model-based monitoring of uncertain systems. Considering its application in the domain of flight control, the resulting monitor requires no physical modelling of the aircraft and allows to account for the effect of additional lag in the closed-loop’s response due to the dynamics of the actuation system. Thus, failure detection capability with respect to allowable system responses derived from high-level requirements is improved. The approaches effectiveness is demonstrated in a simulative use case for runtime response monitoring of requirements from ADS-33E-PRF with a nonlinear model of a piloted vertical take-off and landing aircraft in hover flight. Results of optimization-based worst-case analysis are presented in order to assess the monitor’s proneness to generate false alarms. The improved capability to detect failures in the presence of uncertainty is finally validated through numerical simulations.
Original language | English |
---|---|
Journal | ICAS Proceedings |
State | Published - 2024 |
Event | 34th Congress of the International Council of the Aeronautical Sciences, ICAS 2024 - Florence, Italy Duration: 9 Sep 2024 → 13 Sep 2024 |
Keywords
- Incremental Flight Control
- Model-Based Monitoring
- Runtime Monitoring