vor 1 Jahr

HANSA 02-2017

  • Text
  • Hansaplus
  • Maritime
  • Hansa
  • Jahrgang
  • Marine
  • Shipping
  • Vessels
  • Hamburg
  • Ships
  • Cylinder
  • Schiffstechnik
Suezkanal Economic Zone | Wilhelmshaven | Decksausrüstung | Lubrication | Lärmschutz | German Shipyard Survey | Ship of the year | Searoad Mersey II | Versicherungen | Singapur | NOK

Häfen | Ports

Häfen | Ports Navigation inside ports can be tricky even on a conventional ship. A lot of factors have to be considered for unmanned operation Port environments add complexity Some players in the maritime industry are urgently working on the subject of autonomous shipping. While the concepts might work fine on the high seas, researchers of TU Delft look at the challenges when it comes to navigation in port areas The modern autopilot in aircraft has been around for decades and cars driving themselves are becoming a reality faster than most people would have thought. It is only a matter of (a short) time until autonomous navigation as a concept finds its way into shipping as well. Autonomous sailing for commercial purposes is already the focal point in various research conducted by companies such as Rolls-Royce and DNV- GL, however in these studies, autonomous navigation in the port environment is left untouched. This environment is characterized by a wide variety of manned vessels, where traffc situations are solved by constant communication between crew and port authority. It is a crowded space, and the inertia and momentum of ships imply that there is limited room for error. So we start to question: »How could autonomous, unmanned navigation be made possible in today’s port, taking into account the navigational and communication complexity between vessels and the port authority?« The navigation in ports is of such complexity that to make it manageable it is best to split it up into different parts. For each of those parts, or control levels as we will call them, there could be various solutions differing in degree of autonomous decision making and human involvement. We applied this idea to our case of having a waterborne autonomous container shuttle transferring containers between various terminals in the Port of Rotterdam. In our search for a solution for the navigational complexity in autonomous, unmanned navigation in ports we identified four control levels in the navigational decision making: strategic, tactical, critical and super critical. The full navigational decision making as proposed is summarized in the figure on the next page. The strategic level spans navigational decisions before the vessel leaves the berth. Therefore this level is restricted to preliminary route planning of the autonomous vessel. Possible route concepts could include a liner service, which could see a fixed and (to others) known route at tight departure times to improve predictability and safety. Because the port authority has best overview of vessel traffc flows, it should have significant influence on this route planning. Also a cargo-based planning could be implemented, in which case the vessel sails according to the needs of the terminals based on whether and whereto containers have to be transported. This requires a more dynamic route, possibly based on historical data of vessel traffc flows inside the port. The tactical level starts when the vessel leaves the berth and follows the route determined in the strategic level as far as possible. It predicts vessel behavior based on live traffc information and acts accordingly. Platooning is a possible solution for autonomous navigation, where the autonomous vessel joins a conventional, manned vessel on the same route. That way, the navigational complexity could be solved by the manned vessel. Remote control by an operator in a Shore Control Center (SCC) is a possible solution for the time in between platooning. A more advanced system might be able to anticipate traffc situations on its own. The autonomous vessel then adapts its sailing profile and route to these predictions in order to prevent critical vessel traffc situations. Within a port there are situations in which the autonomous vessel is unable to predict the behavior of others. These situations are part of what is called the critical level. An important aspect of this control level is that communication with the other party to safely resolve the situation is still possible. An operator in a SCC would take over some of the control, although the role of the autonomous vessel can still be considerably large, e.g. the autonomous vessel could propose a solution to the operator who in turn can approve or deny. The final control level is called super critical, which can be seen as an extension of the critical level as it also deals with unexpected situations. However, now communicating with the other party to deal with the situation is not possible. Examples of situations that classify as super critical are the encounter with a person or object in the water, or the confrontation with unexpected behavior of vessels with whom no communication is possible (for instance due to a black-out). Remote surveillance, although introducing reduced autonomy of the vessel and requiring high concentration and workload for the operator, or a blacklist could stop the autonomous vessel in the case of such danger as a first response. An autonomous vessel would use a 84 HANSA International Maritime Journal – 154. Jahrgang – 2017 – Nr. 2

Häfen | Ports Photo: Felix Selzer Unsere Häfen. Ihre Zukunft. blacklist to identify hazardous situations and act corresponding to predefined reactions. The SCC would (partly) take over the control. Most of the technology required for sailing autonomously is already available. Important lessons regarding sensors, actuation mechanisms and control algorithms can be learned from the car and aerospace industries. When self-learning artificial intelligence is brought into the picture, we might need to look elsewhere, but that leads us onto the next remark. It is still expected that a human should be kept in the loop for handling situations where other vessels behave unpredictably. A combination of autonomous detection of unexpected situations and human correction could be a perceivable first step. In that way public acceptance of the autonomous system could also be increased before entirely autonomous behavior is introduced. Further development of currently used systems such as AIS and ECDIS used on most vessels sailing in ports can benefit the implementation of autonomous shuttles greatly by incorporating route-sharing features and more extensive ship and cargo information. Finally, the effectiveness of the implementation of an autonomous shuttle for the purpose we envision is greatly influenced by the cooperation of multiple parties like the port authority and terminal operators. Autonomous navigation of vessels in ports has not previously been addressed in research, due to its complexity. By dividing autonomous decision making into four control levels, the complexity will become manageable and the concept of autonomous navigation is brought closer to reality. This does not take away, however, that critical traffc situations remain diffcult to handle. After autonomously identifying an anomaly, the control should be handed over to a shore-based operator. Data sharing to improve predictability of the autonomous vessel’s behavior is also recommended and benefits the implementation of the autonomous shuttle concept (the full paper containing can be found on Authors: Maurits van den Boogaard, Andreas Feys, Mike Overbeek, Joan le Poole, Robert Hekkenberg, TU Delft, HANSA International Maritime Journal – 154. Jahrgang – 2017 – Nr. 2 85

HANSA Magazine

HANSA Magazine

Hansa News Headlines