Toward a Distributed Case-Based Reasoning for Human-Machine State Diagnosis: Application to Railway Driver State Analysis

This paper purposes a specification of distributed case-based reasoning for human-machine state data-based diagnosis. Data is acquired using non-intrusive sensors such as eye-trackers or cameras aimed at the human operator face. The proposed distributed case-based reasoning formalism allows different configurations. In real-life applications, some sensors are unable to deliver correct information in particular situations, such as low light environments or when the human operator face gets outside the camera field of vision. The proposed approach can deal effectively with such loss of information.