As the desire to equip Autonomous Surface Vessels with increasingly complex autonomy grows, development of effective testing methodologies for autonomy-based behaviors is of importance. Traditional simulation-based single event testing, however, lacks the ability to evaluate the performance of autonomy algorithms before in-situ testing takes place. One issue with designing events in long-run simulations with multiple consecutive events is that the state of the system under test is not known before runtime. Therefore, methods for designing interactions that are specific, yet versatile and flexible are required. This paper presents a method to deterministically synthesize maritime traffic interactions that can be presented to a system under test regardless of the state of the SUT. The development was motivated by three factors. First, the developed method needs to allow a targeted interaction design. Second, the method should accommodate closed-feedback testing approaches that can select testing situations based on prior performance of the SUT. Finally, the method must facilitate testing time compression by reducing and/or eliminating time during the simulation when the SUT is not being stressed by external factors. The method was validated via a simulation. Various scenarios were designed, both trivial and non-trivial and the resulting interaction data was recorded. The overall approach and results presented here validate the use of this method to synthesize varying intensity maritime traffic interactions. Results indicate that the approach can enable more robust evaluation of maritime autonomous algorithms.
Keywords
EVALUATION,SCENARIOS,SIMULATIONS,TESTING
Additional Keywords