Site planning and on-board collision avoidance software to optimize autonomous surface craft surveys

Autonomous surface craft (ASC) have strong potential as platforms for repeat transect surveys to meet oceanographic sampling needs in coastal and estuarine systems. Good spatial and temporal coverage and resolution could be achieved by an ASC capable of weeks-long operations at an average speed of 5 knots, the design goal for the Surveying Coastal Ocean Autonomous Profiler (SCOAP) catamaran ASC currently under development. This project addresses remaining operational challenges for long duration ASC deployments: site planning, and on-board collision avoidance (CA) software, to help minimize risk of collisions with other vessels. To help plan an anticipated SCOAP deployment in Rhode Island Sound (RIS), a method to use archived Automatic Identification System (AIS) data for site planning was developed. 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 hypothetical repeat transect ASC surveys under consideration based on desired oceanographic sampling goals. Average vessel track frequency varied from as low as 0.01 tracks per day in the relatively untraveled areas of southeastern RIS to about 8 tracks per day in the busy shipping lanes near Narragansett Bay. The corresponding number of potential collisions for a month-long repeat transect ASC survey in these two areas was about 0-2 and 12-22, depending on the time of year, respectively. Based on experience in the RIS area it is expected that 40-70% of vessels will not have an AIS; therefore the total number of encounters is expected to be appropriately higher. The analysis demonstrates a quantitative means of site planning that balances oceanographic sampling needs against associated operational risks.;On-board autonomy software capable of performing collision avoidance (CA) maneuvers will be essential to long duration ASC operations. Simulations of repeat transect ASC surveys in the presence of traffic vessels were carried out using the Mission Oriented Operation Suite (MOOS) Interval Programming (IvP) Helm autonomy software, which will be implemented on SCOAP, in order to demonstrate its CA performance. Non COLREGS based behaviors were used within the MOOS IvP Helm framework. The ASC CA maneuvers fell into a small number of distinctive encounters, a large deflection, course-reversal, and leave/return to station keeping. Sensitivity to station-keeping time, vessel speed, avoidance radius, traffic vessel angle, and traffic collision avoidance (TCA) was investigated. TCA reduced the magnitude of the ASC course alterations while dependence on other parameters was modest. A multivehicle encounter was tested and the algorithms proved capable of resolving multi vessel conflicts. Overall the CA algorithm performed reliably and demonstrated that it is suitable for application to repeat transect surveys, resolving traffic vessel encounters efficiently with minimal disruption to the oceanographic sampling goals.;The US Coast Guard (USCG) Collision Avoidance Regulations (COLREGS) are a set of rules, both prescribed navigation behaviors and explicit exceptions to them, governing manned vessels. Because applicability of COLREGS to ASCs is currently unresolved, a framework was developed to guide the best manner of approaching COLREGS compliance for different ASCs in various applications. The planned transit and/or station-keeping of the ASC and its maneuverability relative to typical traffic vessels are considered. In the case of a repeat transect survey, the best approach is a mixture of exceptions and prescribed behaviors: to always give way, except when station keeping and limited in maneuverability, for example due to in water sensors. Based on the framework it is recommended that CA capabilities of on-board autonomy systems be designed with flexibility to implement different combinations of prescribed behaviors and exceptions for different ASCs and applications.