Collision avoidance is one of the most crucial tasks when operating a ship. To prevent the collision, the ship must be provided with sufficient information, including the current position of the own ship and the current and future position of any proximate ship. With the help of cutting-edge sensor technologies, the current position of the own ship and nearby obstacles can be accurately estimated. However, predicting the future position or intention of any proximate ship takes more action.
Conventionally, when two manned crewed ships encounter, they can exchange or negotiate their intention on how to navigate next by easily using verbal communication through a radio connection. However, this method cannot be used for autonomous systems due to the “language barrier” between humans and machines. One alternative solution is to assume that the proximate ship won’t change its intention and keep the current speed and course. This method is effective for seagoing ships since they often do not change their course and speed in a short period. Nevertheless, when it comes to inland waterways, where the traffic density is higher, this assumption does not hold. A possible way to overcome the aforementioned struggle is to use an intention model to predict the possible change in the course and speed of neighboring ships. Accordingly, by considering different underlying factors, e.g., a collision risk of the proximate ship with another obstacle, the intention model could predict the future behavior of the neighbor ship and offers a proactive approach for collision avoidance [1].
Despite the merit of the intention prediction method, there is still a small probability that the prediction could fail. Besides, considering that cutting-edge telecommunication technologies allow ships to send and receive a large amount of information in a short time, we can exploit this communication ability to exchange/negotiate intentions between ships automatically. With the help of information exchange, several ships can collaborate to solve the collision avoidance task. Eventually, by collaborative solving the collision avoidance task, not only the benefits of the own ship, e.g., energy saving and navigational safety, are guaranteed but also that of neighboring ships. Indeed, a negotiation framework for collaborative collision avoidance has shown its potential to guarantee safe navigation for ships in inland waterways [2].
When it comes to information exchange between ships, there are problems that need to be solved before they can be used in traffic. Firstly, the current research primarily focuses on the communication protocol between machines, which means that manned ships cannot participate in this kind of communication. As future traffic scenarios will expect mixed traffic, including both manned and uncrewed ships, a protocol designed for manned only or unmanned only is not practical. Secondly, since the intention of each ship is available within the communication network, it becomes a potation target for cyber-attack. Therefore, protecting the integrity and guaranteeing a secure network for communication is also essential.
Developing a communication protocol that provides the necessary conditions for collaborative operating between ships (both manned and unmanned) is challenging. However, it opens opportunities to apply advancing techniques in control theory to solve this problem. Through the AUTOBarge project and my PhD project, I hope my work can contribute to solving this problem.
An article by Hoang Anh Tran
References:
[1] Rothmund, Sverre Velten, et al. “Intention modeling and inference for autonomous collision avoidance at sea.” Ocean Engineering 266 (2022): 113080.
[2] Chen, Linying, et al. “Cooperative multi-vessel systems in urban waterway networks.” IEEE Transactions on Intelligent Transportation Systems 21.8 (2019): 3294-3307.