This patent introduces a technology for adaptation of one or more system variables used for vehicle longitudinal and lateral motion control.
The adaptation is done online during vehicle operation using a General Regression Neural Network at its heart. In one extent the system can for instance estimate parameters of a kinematic single-track model to increase control accuracy.
The Parameter adaptation can furthermore use information about future target set points of vehicle states to create predictive control adjustments to compensate system dead times.
Source: Technology for model parameter adaptation of a dynamic model for lateral and longitudinal guidance of a motor vehicle (direct link to pdf)