Data-Driven Approach for Feature Drift Detection in Embedded Electronic Devices

Oussama Djedidi (1, 2) , Mohand Djeziri (3, 1, 2) , Nacer M'Sirdi (2, 1)

1 SASV - Systèmes Automatisés à Structure Variable
2 LIS - Laboratoire d'Informatique et des Systèmes (LIS) (Marseille, Toulon)
3 LSIS - Laboratoire des Sciences de l'Information et des Systèmes

This paper is a part of a project aiming to develop supervisor and monitoring devices for embedded systems in airplanes and vehicles. It focuses on the reliability of these systems and establishes a monitoring framework to detect drifts and faults in the behavior of the heterogeneous central processing units (CPU) and graphics processing units (GPU) chips powering them. In this work, we use a previously developed incremental model of these chips and associate it with a fault detection algorithm. Estimations from the model constitute inputs to the diagnosis module. The latter generates alarms in the presence of faults or drifts in the characteristics and features of the System-on-Chip (SoC). The obtained results validate the proposed monitoring algorithm and demonstrate the effectiveness of the fault detection algorithm.

Mots clés: Analytical redundancy; Embedded systems; Modeling; Monitoring; Microprocessors; System identification; System-on-chip

Domaines: Automatique / Robotique; Traitement du signal et de l'image [eess.SP]; Electronique

Oussama Djedidi, Mohand Djeziri, Nacer M'Sirdi. Data-Driven Approach for Feature Drift Detection in Embedded Electronic Devices. 10th IFAC Symposium on Fault Detection, Supervision and Safety for Technical Processes (SAFEPROCESS 2018), Aug 2018, Varsovie, Poland. ⟨hal-01869747v1⟩