Design of a lane-tracking driver steering assist system and its interaction with a two-point visual driver model

In this paper we investigate the design of a lane-tracking driver-assist system and, in particular, its interaction with the well-known two-point visual perception driver model. We present two different driver-vehicle interconnections and two different control designs that assist the driver in the lane-tracking task; the first design is based on linear multivariable output regulator theory, and the second design is based on model predictive control (MPC) theory. In the first interconnection the controller is “blind” to the driver's actions, and an appropriate blending of the controller and driver actions is used in order to steer the vehicle. In the second interconnection the control design takes into consideration the driver model, and the control steering command is shared between the lane-tracking steering assist system and the driver. The performance of both controllers and interconnections is compared and evaluated using three different drivers and also against a unassisted driver-only scenario.