An adaptive nonlinear model predictive controller for longitudinal motion of automated vehicles

This paper presents an Adaptive Nonlinear Model Predictive Controller for longitudinal motion of automated vehicles which incorporates advance information on future speed demand values, as well as on road grade changes. It is used in combination with a state and parameter estimator to adapt to a changing vehicle mass. This allows improved speed tracking capability for horizontal driving and steep hill climbing. Performance is explored through simulation of a driving scenario in a parking garage and shows encouraging improvements in quality of control. It could be shown that even though the controller is optimized for tracking performance, average fuel consumption can be reduced.