Robust Preview-Based Tractor-Trailer Lateral Control For Lane Keeping

This work proposes a control framework for automated lane keeping of a tractor-trailer truck. The proposed architecture controls vehicle variables in a preview-based manner, thus gaining robustness to sensing and acting delays/lags. Electronic power steering of tractor front wheels is used to act on lateral and yaw dynamics, as well as vision-based lane mark detection for loop closing. A feedforward/feedback control approach is used to stabilize the truck dynamics formulated as a Linear Parameter Varying system. The design objectives include reference yaw rate tracking, robustness against truck modeling uncertainties and control effort definition. Validation results on a high fidelity simulation environment demonstrate the algorithm performance both in constant and varying speeds scenarios, as well as its robustness to external and internal disturbances.