Collaborative collision avoidance methods for autonomous ship navigation
Abstract
The rise of autonomous shipping promises to revolutionize maritime transportation. As autonomous ships increasingly integrate into existing maritime traffic, the development of robust collision avoidance methods becomes paramount. Whereas current collision avoidance techniques in autonomous ship applications heavily rely on onboard sensors, communication and information exchange between autonomous ships and other vessels can reduce uncertainty in the operating environments and improve collision avoidance decision-making. This thesis presents methods to implement collaborative strategies in autonomous ship collision avoidance applications.
The thesis focuses on four key maritime actors who interact and participate in various collaboration scenarios. These actors are identified as autonomous ships, conventional vessels, Vessel Traffic Services (VTS), and Remote Control Centers (RCC), and are explained in the following paragraphs.
First, despite advancements in autonomous technologies, conventional vessels operated by seafarers will remain a staple of maritime traffic for the foreseeable future. Therefore, autonomous ships must be able to coexist and safely navigate alongside these conventional vessels.
Second, the increasing number of autonomous ships will lead to a diverse maritime environment with vessels possessing varying capabilities. This necessitates autonomous ships having robust mechanisms for interaction, information sharing, and conflict resolution – not only with conventional vessels but also with each other.
Third, designated VTS manages maritime traffic in busy waterways such as ports, straits, canals, or congested waters. The VTS plays a crucial role by monitoring traffic, interacting with individual vessels, and providing them with vital situational awareness information and recommendations related to the traffic and specific regions. As autonomous ships integrate into these environments, a similar level of information exchange and collaboration with VTS will be essential. This collaboration will enable VTS to provide targeted support to autonomous ships, enhancing overall safety and efficiency within these complex traffic zones.
Finally, RCCs play a critical role in collaborative collision avoidance strategies for autonomous shipping. The concept of autonomous shipping encompasses a spectrum of operational autonomy levels ranging from remote-controlled to fully autonomous. RCCs provide essential human oversight for all levels of autonomy. Even fully autonomous voyages may require intervention in areas with heavy traffic or complex navigational challenges. RCCs fill this gap by allowing shore-based personnel to monitor these situations and intervene when necessary. This ensures that a human element remains present in the system, ready to make critical decisions and prevent potential collisions.
This thesis presents a series of novel contributions towards collaborative collision avoidance methods involving these four key maritime actors. These contributions are presented in detail across dedicated chapters, drawing upon research findings documented in articles submitted or published in peer-reviewed journals. The main contributions can be summarized as follows.
The first main contribution lies in a comprehensive literature review that lays the foundation for the subsequent research. This chapter provides a critical examination of current and future technologies and concepts with the potential to facilitate collaboration between the various maritime actors. It offers an extensive overview of communication system alternatives and presents a comparative analysis of previously conducted research in the field of collaborative collision avoidance for autonomous ships. By establishing a thorough understanding of the existing landscape, this literature review paves the way for the development of novel and effective collision avoidance strategies explored during the PhD.
As a second major contribution, the thesis proposes a novel ship-to-ship route exchange-based collision avoidance method. This method emphasizes the importance of collaboration between autonomous and conventional ships by facilitating the exchange of trajectory plans. By sharing this critical information, vessels can make informed decisions and proactively manage potential conflicts. The proposed method leverages a hierarchical collision avoidance architecture, enabling vessels to agree on safe, rule-compliant, and collision-free plans at greater distances. Furthermore, a specifically designed optimization algorithm prevents close encounter situations, minimizing the need for risky last-minute maneuvers. Recognizing the importance of real-world application, the method considers adherence to the International Regulations for Preventing Collisions at Sea (COLREG) and incorporates mechanisms to handle non-cooperative vessels, ensuring safety even in unpredictable scenarios.
Collaborative strategies for collision avoidance involving both conventional and autonomous ships necessitate different approaches compared to scenarios where only autonomous vessels are present. As a third major contribution, the thesis investigates the scenarios where multiple autonomous ships collaborate and introduce a negotiation-based collaborative collision avoidance framework. This approach tackles the challenge of diverse decision-making processes by enabling ship-to-ship intention exchange. Through this exchange, autonomous vessels—even those equipped with different collision avoidance algorithms and safety parameters—can reach a consensus on safe and rule-compliant maneuvers. The framework empowers autonomous ships to propose both egocentric (self-preserving) and altruistic (considering others) behaviors based on their individual risk tolerance and obligations under maritime traffic rules. Individual autonomous ship collision avoidance algorithms adhere to the main COLREG rules related to navigation and the framework is tested with non-cooperative vessels in the environment. The effectiveness of this method is further substantiated through real-world field experiments, demonstrating its practicality and potential for real-world implementation.
The fourth major contribution of this thesis delves into the potential for enhanced collaboration between VTS and autonomous ships. This section explores the possibility of future VTS centers assuming greater authority and direct responsibility for collision avoidance planning involving autonomous vessels. This approach envisions a central VTS unit that collects information from all participating vessels within its designated area. By leveraging this comprehensive data set, the VTS centrally solves the overall collision avoidance problem. Subsequently, the VTS translates these solutions into actionable waypoints and speed plan messages, which are delegated then back to the individual vessels within the environment. As in the previous contributions, this method also considers the main COLREG rules and proposes mechanisms to handle non-cooperative vessels.
The final major contribution of this thesis proposes a decision support system for RCC operators. This system investigates the collaborative approach of integrating autonomous navigation systems with human expertise for trajectory planning. This approach leverages the computational power of autonomous technology while capitalizing on human experience in multi-criteria decisionmaking. First, the decision support system incorporates environmental factors such as ocean currents, wind, and tidal information, along with considerations for narrow channel geometry, squat effects, and under-keel clearance concepts. Then, the system utilizes these factors and integrates vessel maneuvering characteristics and potential failure scenarios to conduct a comprehensive risk assessment. Next, the optimization algorithm calculates numerous alternative trajectory plans. Finally, human expertise and preferences are integrated through multi-criteria decision-making (MCDM) and clustering methods to filter and present a manageable set of alternative solutions to the RCC operator for final decision-making. This collaborative approach ensures that human oversight remains an essential element in navigating complex situations.
In conclusion, this thesis investigates various collaboration scenarios and strategies applicable to autonomous ship collision avoidance methods. Through its contributions, this thesis aims to significantly enhance the safety and efficiency of maritime navigation in the era of autonomous shipping.