Vehicle and driver modeling and threat assessment for driving support functions

The article reports a novel method to assess the driving risk level and design a human friendly warning strategy. The method is built on a Receding Horizon (RH) approach that is instanced for a set of predefined driving scenarios such as driving in the lane, change lane, etc. In control field, the RH is a technique that solves a sequence of optimization problemin real-time and, at each time step, applies only the first value of the control plan to steer the system towards a desired behavior. In this work, differentlythan in the control application, the initial value of the each control plan is used as a measure of the correction that the rider should apply to conform to the computed optimal maneuver. This choice has the advantage to provide an homogenous measure of the threat independently from the scenario and it is directly linked with the control variable that the rider should use to accordingly changethevehicledynamics. Additionally,theRH approachnaturallyaccommodatesroadgeometry and attribute constraints, vehicle dynamics, driving input and styles. A proper development of the vehicle model and a quantitative characterization of the human driving skills play an important role in the method effectiveness. Additionally the method make use of a dedicated solver to compute the probleminrealtime. Themethodwas appliedwithsuccess todevelopdrivingsupportfunctionsboth for cars in the the FP6th European project PReVENT and the FP7th interactIVe and for motorcycles in the FP7th European Project SAFERIDER. The article introduces the RH approach as defined for the driving threat assessment. Then it discusses in details the vehicle modelling requirements and how human driving skills are included in the proposed method. Examplary use of how the system works in different driving scenario will be given. Finally, the experimental results of pilot tests are shown for all the developed applications.