Page - (000009) - in Integration of Advanced Driver Assistance Systems on Full-Vehicle Level - Parametrization of an Adaptive Cruise Control System Based on Test Drives

Abstract Advanced Driver Assistance Systems (ADAS) support drivers in fulfilling their driving task by reducing workload and enabling a more safe and comfortable drive. However, the increasing market penetration of ADAS, along with the wide variety of types and models, has led to a need for a cost and time-efficient way to integrate and parametrize new systems. One essential point for the integration process of comfort-oriented ADAS is the question of driver satisfaction with respect to safety, reliability, trust and comfort. The current work offers a method for parametrizing an ADAS controller with the help of test drives with non-professional drivers. The proposed method is validated by the parametrizationofanAdaptive Cruise Control (ACC)system, whichsupports thedriver by keeping a desired vehicle speed or defined distance to a proceeding slower moving vehicle, the Object to Follow (OTF). For the selection of the OTF, the prediction of the future path of the own vehicle (ego vehicle) is an essential part of the ACC system. Today, different algorithms are implemented for path prediction. To evaluate these algorithms, test drives were carried out with a specially equipped vehiclewithnon-professional testdrivers. Basedonthemeasureddata,differentmethods forpathpredictionwere compared. Anovel steeringpredictionalgorithmwasdeveloped, which is used in combination with a linear Single-Track Model (STM) to predict the ego vehicle’s path. Based on the predicted ego vehicle path, the OTF is selected, which is then used to parametrize a novel ACC controller. The performance of the controller fulfils previously defined safety and comfort requirements, as well as string stability. SimulationswiththerecordedOTFdataas inputwerecarriedout,whichshowedthatthe ACC controller is able to simulate the behaviour of the human driver. Furthermore, the controller cuts acceleration peaks, which leads to a more comfortable feeling than with the measurements obtained when the human drove the vehicle. Finally, a comparison withmeasurementsofa state-of-the-artACCsystemshowedsimilarbehaviourcompared to the production controller in following another vehicle. The results of the present study show that the proposed method is able to identify an appropriate set of parameters for an ACC controller. The idea of parametrizing the con- trollers with the help of human test driver should lead to a human-like behaviour and increase customer acceptance of the system. Additionally, this optimized parametriza- tion method will help to shorten the development and validation process, which is very important for saving costs.