Micro-slip control of automated manual transmission systems through Hybrid Model Predictive Control methodologies

The increasing needs of car drivers in terms of fuel economy and comfort have pushed the automotive industries towards the development of innovative transmission systems and control methods to guarantee better performance and high efficiency in modern vehicles. A suitable transmission of the internal combustion engine torque, which is non-uniform and oscillatory by its nature, is crucial in this perspective, since it can lead to high mechanical stress on driveline components and unpleasant feeling on car passengers. Micro-slip control approach for Automated Manual Transmission (AMT) is a well affirmed technique among car manufactures thanks to its effectiveness in the attenuation of vibrations and smooth shifting. In particular, the micro-slip strategy aims at controlling the transmitted torque to avoid the complete engagement of the clutch by regulating the relative speed between the clutch discs. In this context, a problem of paramount importance is the sign inversion of the clutch relative speed that occurs when sudden acceleration/deceleration maneuvers are performed. In such a situation, the dynamical model of the transmission changes making challenging the design of the micro-slip controller. In the present thesis, developed in collaboration with Centro Ricerche Fiat (CRF), an original approach to micro-slip control based on Model Predictive Control methodologies is introduced aimed at attenuating the torsional oscillation in the transmission driveline and accounting for the sign inversion in the clutch relative speed. The first contribution of the thesis is the development of a suitable Piecewise Affine (PWA) model of the transmission system that explicitly describes the sign inversion phenomenon of the clutch relative speed. The next contribution is the introduction of an original formulation of the micro-slip control problem in the framework of a Hybrid Model Predictive Control that can effectively handle the design problem in the presence of the considered PWA model. Finally, extensive simulation tests performed on a realistic non-linear transmission model are introduced to show the effectiveness of the proposed approach.