Driver assistance mechatronic systems are standard features of the modern car. This is particularly true of systems that use braking systems (ABS, ESC, etc.). Mechatronic systems directly related to steering mechanisms are very limited, as yet. They apply mainly to support the driver’s effort (servo-type systems), and to stabilize the vehicle’s trajectory (active steering systems, 4WS systems), in which powering and gearing are speed dependent. Full automation of driving includes, as of now, only the parking maneuver (Parc Assist System), which is implemented at a very low speed - at quasi-static conditions. Full automation of road maneuvers at high speeds (when the car should be treated as a dynamical system) remains difficult and is still open. Automatic control of lane change is a key to automate more complex maneuvers (eg. avoiding, overtaking etc.).

The subject of automation of lane change was undertaken by the authors in the research project on the control of the vehicle with suddenly appearing obstacle. The authors’ model of a conceptual control system was presented at the Conference ESV'2015. The aim of extensive simulation studies was:​ testing of the controller operations, and evaluation of its sensitivity to changes of the vehicle and road parameters. This paper presents unpublished results of these studies.

The lane change controller has a mixed structure. In the open-loop structure it works as a set-point signal generator which generates three variables (signals) determining the lane change maneuver:​ a set-point input signal of steering system angle, and two set-point output signals describing vehicle’s motion. In the closed-loop structure it works as a steering signal corrector which corrects on-line (by two Kalman regulators) the steering system angle signal. The set-point signals, as well as regulators’ algorithms are based on a simple reference model (simplified "bicycle model"). In simulation, the virtual object of control – the model of medium-duty truck is very detailed (MBS-type, 3D, nonlinear). This model had been verified experimentally.

Due to the complexity of the vehicle motion model, a sensitivity analysis must be based on comparing results obtained from the simulation with nominal and changed models. In order to objectify the analysis, special integral indexes have been introduced. They use the signals from nominal and changed models. The results of simulation show that the proposed concept of the automatic control is good.

The sensitivity study focuses on the variation of parameters that appear to be crucial for the correct operation of the control system. Bearing in mind the experience of drivers and researchers, difficult situations (slippery road, vehicle unloaded, high speed), as well as measurement errors are taken into account in the simulation investigations.

The presented method of automatic control can be an attractive proposition for designers and researchers of active steering systems which enhance active safety of vehicles. The subject of the work is directly related to the subject of the session Enhancing Safety with Connected and Automated Vehicles.