In the development of autonomous systems, the automotive industry has primarily focused on the application on paved roads, which are usually measured precisely in advance to create HD maps. For this purpose, a large number of complex sensors are used to enable the vehicle to perceive the traffic environment. In the military area, however, there are severe restrictions with regard to the usable sensors (non-emitting), and the main area of application is the semi-paved or unpaved terrain.
The development of localization and navigation software of autonomous vehicles re-quires realistic vehicle input data in different scenarios and environments. To achieve higher flexibility and lower costs, this data should preferably be generated from simulation, so that the systems can be tested as soon as possible without the need for ap-propriate hardware. This enables early identification of possible problem areas and avoids unnecessary development loops. Similarly, individual subsystems can already be verified in the simulation at an early development stage.
The main objective of this project is to develop and demonstrate a simulation environment for the virtual testing of autonomous vehicle systems in semi-paved and un-paved terrain, with the special boundary condition that no emitting sensors are used. The simulation environment will help to virtually test the performance of different vehicle and sensor concepts as well as to develop optimized trajectory and maneuver planning for real-world use.
The following research questions are addressed:
1. What methods and tools are suitable for simulating autonomous vehicles and convoys on semi-paved and unpaved roads and paths?
2. Which scenarios and variants need to be simulated to support the development of localization and navigation software of autonomous vehicles?
3. How can the simulated models be integrated into vehicles to design a flexible and real-time system architecture for passive sensors?
4. When using non-emitting sensors only, is the performance of environmental detection and trajectory planning sufficient to ensure safe autonomous convoy operation?