Page - 42 - in Integration of Advanced Driver Assistance Systems on Full-Vehicle Level - Parametrization of an Adaptive Cruise Control System Based on Test Drives

3. Development Process max,setτvehicleA, min,setτvehicleA, min,setτvehicle B, 3020100 130 110 90 70 [s]t Figure 3.2.: Measurements for manoeuvre “Approach on slower moving target in the same lane”, adapted from [Hol12] and [BHLSE13] This huge database is used to train neural networks. To validate the tests, data that is fed into the neural network has to be recorded. The output should be the same as if probands asses the system. It is a promising approach because they have already used it extensively to evaluate the driveability of drive trains, and it has performed quite well. Figure3.2showsanexample forthemanoeuvrecalled“Approachonslowermovingtarget in the same lane”. Here, the ego vehicle is driving at a constant speed of vvx= 130km/h and isapproachingatargetwithaspeedof 0vT = 80km/h. Themanoeuvre is conducted three times, two times with vehicle A and one time with vehicle B. For vehicle A, the time gap is set to τset,min= 1s and τset,max= 2.3s. The measurement with vehicle B is made with the setting τset,min= 1s. It is evident that the behaviours in this manoeuvre differ significantly, even between the two measurements of vehicle A. Some of the test drivers rate vehicle A better than vehicle B, although vehicle B has less undershoot in speed, [BHLSE13]. The different evaluations demonstrate why an objective assessment is very difficult. From the point of view of control theory, less undershoot is better, but drivers may assess it in another way. 3.2. System and Component Levels Forthedevelopmentandvalidationonthe system andcomponent levels, simulationtools, module and component test benches are used. Often, a module test bench is used for the validation and development of camera-based ADAS. In this case, a camera faces a monitorwhere the Field of View (FOV)of the sensor is simulated. Suchasetup is called Hardware-in-the-Loop (HIL).Thepictureon themonitor is ananimationof a simulation 42