FDIR

Our main research activities in FDIR sector are:

• Development of a Model-Based module based on the Fuzzy Inductive Reasoning (FIR) together with the Envelope technique to manage the decision making process on symptoms and to output gravity index level, with different tasks and benefits.
• Implementation of a Knowledge-Based module which takes advantage of a Possibilistic Abductive algorithm. It works on a Knowledge Base (KB) representative of a map of uncertain causal dependencies between failures (or anomalies) and their effects (symptoms) exploiting the Possibility Logic as uncertainty handling formalism.
• Definition of a standard approach for Recovery using innovative Automated Reasoning (AR) and Decision Making (DM) techniques.
• Implementation of the studied methodologies in a virtual environment and then on a physical robot.
• Characterization of efficient paradigm for interaction among different agents in the framework of Distributed Recovery.

One important result of this research is the SMART-FDIR project realized as a sub-contractor of Alenia Spazio (Turin - IT) in the framework of an ESA activity: we validated a very innovative approach to FDIR on the models of the power and AOCS subsystems of the GOCE spacecraft.

Detection of a failure in the mobility system with the FIR technique