Page - 56 - in Integration of Advanced Driver Assistance Systems on Full-Vehicle Level - Parametrization of an Adaptive Cruise Control System Based on Test Drives

5. Selection of the Object to Follow by the camera mounted in the vehicle. They demonstrated that this method even works at high vehicle speeds. Anotherapproach is touseadditional environmental-recognition sensors, suchas camera systems, to detect the lane markings on the road, [WDS09]. This works properly if the vehicle is not changing lanes and if simple situations concerning the lane markings occur. For camera systems, it is very difficult to detect which lane the vehicle is driving in when more potential possible solutions can be found (e.g. at road construction zones or at road intersections). To handle such situations, the work of Yim et al. could also be used, [YSZS10]. In general, direct measurement of the lane markings places high demands on the accuracy of the optical system. Additionally, detecting the ego vehicle driver’s intention to change lanes by sensing the indicator usage will help to increase the quality of the predicted path. This could also be done for vehicles travelling in front of the ego vehicle, and the data could be sent to the ego vehicle via V2V technologies or by analysing the video data recorded by the camera. The ability to detect the front vehicle’s intention to change lanes via camera strongly depends on how often the front vehicle’s turn signalflashes. The researchofFro¨hlichetal. in [FEF14] showedthat if the turn signals flashes at least three times, the system generates very reliable information, with a very low number of wrong detections. In modern vehicles, the indicator flashes three times even if it is activated by the driver for a very short time. This function is called one-touch indicator and helps to fulfil the requirements of Fro¨hlich et al.. These are all promising technologies, [WDS09], but for the current investigations, they were excluded to reduce the complexity of the system. In this chapter, the described algorithmsarebasedonvehicledynamicsdata,which isavailable ineverymodernvehicle equipped with an Electronic Stability Control (ESC) system. 5.1.1. Path Prediction Using Constant Curvature Hypothesis The Constant Curvature Hypothesis predicts the vehicle path based on the assumption that the corner radius the vehicle is driving will remain constant in the future, [WDS09]. For the present model, the radius is not used because the radius for straight driving runs towards infinity. Instead, the reciprocal value, called curvature, κ= 1R is used. The radius R is the distance between the vehicle’s Center of Gravity (CG) and the Instantaneous Centre of Rotation (ICR), see fig. 5.1. The following paragraphs describe how to estimate the actual curvature. All algorithms use the assumption that only small lateral velocities occur vvy 1m/s, resulting in vvx≈ vv. In [WDS09], Winner et al. showed that the curvature can be estimated by κωz = vωz vvx (5.1) at higher vehicle speeds. AnotherapproachdescribedbyWinner in [WDS09]basedonthe lateral acceleration vay 56