A hybrid observer approach for driver drowsiness detection

Abstract

This paper discusses an algorithm for driver drowsiness detection based on the theory of hybrid observers. A hybrid observer is designed on the basis of a driver's model whose parameters change according to the assumed drowsiness levels which are identified on-line from measurements of relevant vehicle signals. Simulations have been undertaken with Carsim to verify the capability of the proposed methodology to correctly identify on-line the current driver's drowsiness level.