• Estimating driver performance from steering behavior. • Observer-based methods for dedicating and reconstructing time-varying driver faults. • Adaptive shared control scheduled by real-time driver performance monitoring • Improving Lane keeping through linear parameter-varying control function of vehicle speed and driver performance. • Simulations and experiments results validate the proposed shared-control strategy. This paper introduces a fault-adaptive shared control approach for driver assistance in semi-automated vehicles. The difference between an ideal and an actual driver is estimated using two methods: a descriptor H ∞ filter and a cascaded Luenberger observer with a linear filter. Unlike related works, the proposed estimation strategies accurately reconstruct the time-varying deviations, considered as a fault, without using restrictive assumptions. The estimated driver deviation is then used in a linear parameter-varying controller scheduled by longitudinal velocity and driver performance, improving steering assistance and overall tracking accuracy. The proposed methods are validated using simulations and in an experimental scaled platform.
Hammoud et al. (Tue,) studied this question.