Criteria and rule based obstacle avoidance for USVs
Aaron Hao Tan, Wong Chee Wee, Timothy Joe Tan · 2010
In this paper, an obstacle avoidance solution based on rules and criteria is presented. The solution has two parts. In the first part, a set of candidate maneuvers is generated to track a given path. Each candidate differs in speed and in lateral offset to the tracked path. Thus, the USV has the option to change speeds or switch "lanes", or do both. In the second part, the most appropriate trajectory is selected for execution. High priority objectives, such as safety clearance, are represented quantitatively as rules. Candidate maneuvers that flout any of these rules are immediately rejected. Lower priority objectives, such as minimizing travel time, are represented as a priority list of selection criteria. From the remaining candidates, an iterative process of short listing maneuvers is performed. At each step, a single criterion is used to rank and select the top trajectories. The process is repeated for criteria of decreasing priority until only one maneuvers remains. The presented solution has been implemented and tested in simulation. Some of these simulation results will be presented.