Date of Award
Master of Science in Ocean Engineering
Daniel L. Codiga
Autonomous surface crafts (ASCs) have strong potential as platforms for repeat transect oceanographic surveys in coastal and estuarine systems. Good spatial and temporal coverage and resolution could be achieved by an ASC capable of weekslong operations at average speed 5 knots, which is the design goal for the Surveying Coastal Ocean Autonomous Profiler (SCOAP) catamaran ASC. This project addresses operational challenges for long duration ASC deployments, to help minimize risk of collisions: site planning, and on-board collision avoidance (CA) software.
In anticipation of a Rhode Island Sound (RIS) SCOAP deployment, a site planning method was developed using archived Automatic Identification System (AIS) data. AIS is a real-time, radio-based system for sharing navigation information among vessels. Archived data were used to determine geographic and seasonal patterns of RIS vessel traffic, and associated frequency of potential collision encounters for various hypothetical repeat transect ASC surveys under consideration. Seasonal-mean AIS vessel traffic varied from as low as 0.01 tracks per day or less in southeastern RIS to about 8 tracks per day in shipping lanes near Narragansett Bay. Corresponding numbers of potential collision encounters for a month-long repeat transect ASC survey in these two areas are about 0-2 and 12-22, respectively, with lower/higher values in winter/summer. Crudely estimated non-AIS traffic suggests up to 3-4 times higher total (AIS and non-AIS vessels) potential encounters. The method provides quantitative information to enable site planning that best balances oceanographic sampling goals against associated collision risks.
On-board autonomy software capable of performing collision avoidance (CA) maneuvers is essential for long duration ASC operations. To demonstrate and evaluate its CA performance the Mission Oriented Operation Suite Interval Programming Helm autonomy software, to be implemented on SCOAP, was used to simulate repeat transect ASC surveys with traffic vessels. The CA algorithm used avoids collisions without attempting to comply with Coast Guard (CG) Collision Avoidance Regulations (COLREGS). Three categories of ASC CA maneuver were identified, in which nearly all encounters fall: large deflection, course-reversal, and leave/return to station keeping. ASC path disruptions are substantially reduced when traffic vessels perform CA in addition to the ASC; sensitivity of encounter statistics to other ASC and traffic vessel path configurations, and CA algorithm parameters, is modest. Overall the CA algorithm performed reliably, resolving at least 97% of encounters in 10s of seconds, including interactions with multiple traffic vehicles simultaneously.
A framework was developed for how ASCs can best approach compliance with COLREGS, the set of rules (prescribed behaviors and exceptions to them) applicable to vessels with human operators on board. The framework considers the ASC operating mode and its maneuverability relative to traffic vessels. For SCOAP, marked for restricted maneuverability and executing repeat transects, the recommendation is to always give way except when station-keeping with sensors overboard. The framework will be reassessed when CG requirements for ASC COLREGS compliance, currently under development, are completed. Regardless, it is suggested that on-board autonomy software include flexibility to implement different combinations of prescribed behaviors and exceptions, for different ASCs and applications.
Filimon, Michael A., "Site Planning and On-Board Collision Avoidance Software to Optimize Autonomous Surface Craft Surveys" (2013). Open Access Master's Theses. Paper 56.