An approach to develop location-based efficiency systems without real test drives

Abstract

The development of intelligent vehicle systems based on non-deterministic input signals of the environment, have a dramatically increased complexity. Real test drives are no longer adequate to cover all test-cases, due to costs and complexity. One way to meet this challenge is a simulated test drive. In this case the driver, the environment and the vehicle are described within simulation models. Scenarios can be reproduced exactly. Tests can be executed automatically. Based on realistic road networks, realistic positioning data can be simulated. This allows development and testing of location-based systems. Especially for electric vehicles location-based systems are of great importance. We developed a simulation framework which represents the three domains (vehicle, driver and environment) within a closed-loop simulation. Using the framework we are able to execute simulated test-drives. In order to confirm the significance of the simulated test drive for the development, we validate the quality of the simulation results with data recorded on an electric vehicle fleet test. The architecture enables to reproduce a real test drive within a closed loop simulation.