A distributed navigation and guidance system for autonomous vessel

Sammanfattning: Using distributed micro-controllers in an autonomous vessel can result in higher robustness and lower energy consumption compared to a traditional centralized approach. However adapting traditional software (e.g. navigation and guidance systems) for a distributed microcontroller system is a complex and challenging task. In this thesis a software solution for controlling autonomous vessels with on board distributed micro-controllers is presented. It is capable of controlling the vehicle by itself, including routing around obstacles in moderately complex environments, or act as an interface to another computer. The routing capabilities come from the use of a high-level path-planner based on RRT, and a low-level vehicle controller based on potential fields to follow the path from the path-planner. By using these methods together the software is capable maneuvering a sailboat between waypoints. In our experiments distributed computing is investigated for the path planning and evaluated in terms of computational time for 1,2 and 3 nodes. A parallelization technique called OR parallelization was tested and it offered a reduction in computational time by 27% with two nodes and 36% for three nodes against a single node. Nevertheless, this gain may not be, significant enough to warrant the extra complexity.