Emulation of Situational Awareness for Autonomous Vessels
Background & motivation
- Autonomous surface vehicles (ASVs) rely on an autonomy system comprised of object detection algorithms, situational awareness and motion planning, see figure below. Due to the complexity of ASVs, simulation-based testing will be an integral part of the verification process.
- The object detection algorithms and the situational awareness depend on a range of electromagnetic radiation sensors (EMRS); e.g. cameras, radar and lidar. Detailed simulation of EMRS require rendering the full 3D environment around the vessel, which is computationally expensive resulting in slow simulations.
- However, the key elements needed to test the motion planning system can usually be simulated much faster, which enables greater test coverage.
- To enable fast simulations of motion planning systems, the thesis should investigate how to emulate an object detection and situational awareness system to provide realistic input to the motion planning system.
- Investigate which effects from situational awareness that are important to emulate for testing performance and robustness of the motion planning system. This can for instance include:
- Noisy obstacle tracks
- Delays in track initialization
- Intermittent tracks
- The student should develop a simulation module in the Open Simulation Platform (OSP) for situational awareness emulation. The simulation module should be integrated with the existing milliAmpere 2 simulator developed by NTNU and Zeabuz.
- The student should develop a list of relevant parameters and failure modes for the situational awareness emulator. This will serve as input when choosing test cases during verification of the collision avoidance system.
- The student will be given access to recorded data from previous experiments. These scenarios should be recreated in the simulator, and the resulting behaviour of the collision avoidance system should be compared to the experiments.
Figure 1: Block schedule of the Zeabuz autonomy system.