4 MAY 2021
The decision-making and logic that is already helping ports reduce collision risks will soon be used onboard to help ship operators plot safe paths through busy waters.
Humankind’s giant leap to the moon is usually summarized as one man’s foot on the lunar surface. When, in fact, it was made possible with many successive small steps. But, like most technological breakthroughs, we tend to omit the details when telling that story.
Shipping’s autonomy story is somewhat similar. Our eyes are glued to the future, when unmanned fleets can sail safely, while crucial developments are taking place today — commercially viable smart solutions for current challenges that also forge the way forward.
“Autonomy isn’t black and white — it’s a series of small steps,” says Alexander Ozersky, Deputy director, Intellectual systems integration, Wärtsilä Voyage. “We’re not trying to make ships fully autonomous tomorrow, but we can retrofit systems that bring new possibilities moving towards less work onboard, less human error and better performance.”
Already today, there are systems onboard ships that can hold a steady course, detect navigational anomalies, optimize routes for weather or fuel consumption, and send accurate estimated times of arrival. And if the ship owners want it, the fully self-sailing vessel will be an evolution of these available systems.
This stepwise approach is the lynchpin of Wärtsilä’s vision for vessel autonomy, detailed in a recent white paper. Since the ability to gauge a situation and make decisions is one of the primary conditions for autonomous navigation, it is also one of the building blocks in Wärtsilä Voyage’s smart autonomy route map.
Developing a vessel’s brain
Moving a vessel from point A to B involves two fundamental steps: One, understanding and interpreting a situation, and two, based on these inputs, deciding the best course of action.
The technology, therefore, involves hardware and software working in tandem: navigation and situational awareness solutions (the senses) such as dynamic positioning system, ECDIS (Electronic Chart Display and Information System), radar and sensors feed data to the intelligent algorithms (the brain) that apply machine learning to interpret the scenario and suggest safe and effective actions.
With new technology and better data exchange, decision-making algorithms have also become increasingly sophisticated, leading to more intelligent systems that have increased autonomous capabilities.
“ECDIS and radar were designed to prevent collisions,” says Ozersky. “Navigation became more advanced with the arrival of ARPA (Automatic Radar Plotting Aid), which could show vessel course and time to collision. Then came a function called ‘trial maneuver’ where you could enter your maneuver, and the system would tell you if it was safe.”
This function — essentially collision avoidance — is part of Wärtsilä’s Advanced Intelligent Maneuvering (AIM) technology. So far, AIM has only been deployed by ports as part of the Vessel Traffic Management (VTM) system. But if AIM could be connected to ECDIS, it could steer the vessel.
For instance, the smart navigation collision avoidance system developed for Singapore’s IntelliTug project was able to take the tug from A to B while dynamically navigating around collision risks. This was made possible with Wärtsilä Voyage’s collision avoidance algorithms that made decisions based on a digital model of the environment created by an advanced situational awareness solution.
Recognizing reality for local maneuvers
AIM can predict vessel movements 30 minutes into the future and suggest maneuvers to avoid collision. This intelligence is based on two primary inputs. First, are obviously the IMO’s traffic codes, or COLREGs (collision regulations), which guides how vessels ‘should’ behave. But these are designed to be human-centric, calling for ‘reasonable’ judgements, for example. So, they are not very computer friendly.
“There are always some ‘local’ driving behaviours to consider. For instance, many of the smaller crafts in Singapore may not comply with COLREGs,” says Ozersky. Hence, the need for the second input — real-life experience of how vessels behave, is crucial for predicting traffic habits and thus guiding local maneuvering.
This ‘real-life experience’ — or more properly, the telemetry data — that feeds into AIM’s machine learning comes from the ports that are already using AIM in their traffic management. Data also comes from vessels that have installed Wärtsilä’s connected ECDIS and intelligent navigation system, Fleet Operation Solution (FOS). One of the FOS modules, Anomaly Detection, logs navigational incidents and recommends corrections. This is all valuable experience to help AIM learn how COLREGs are applied or otherwise in different situations.
The next step towards vessel autonomy is for AIM to be tested for use on vessels. However, before these algorithms can be let loose for physical sea trials, they need extensive digital testing.
Testing waters digitally first
Comprehensive digital testing ensures that the algorithms respond appropriately to any given situation before sea trials can be conducted.
To do this, Wärtsilä Voyage has developed the Steering Control and Autonomy Lab Simulator (SCALab). It is one of the world’s first simulators designed specifically for the testing of autonomous vessels. The SCALab simulator allows users to test the autonomous navigation algorithms using a highly accurate digital twin of the vessel and its sensors within a safe digital environment.
As always, anything digital comes with its inherent advantage — it is much cheaper, faster. It also ensures higher accuracy as a plethora of different permutation and combination of scenarios can be created and tested. This also makes it an enabler for getting regulatory approvals.
“It would have been much more difficult to get approval for IntelliTug if sea trials were our only option,” says Thomas Brightwell, Software Manager, SACA (Situational Awareness Collision Avoidance) at Wärtsilä Voyage. “With simulators, we did not have to wait for regulation to be approved; we could demonstrate the safety of the system to the regulators.”
For the IntelliTug project, SCALab was used to conduct batch testing of thousands of test cases, some of which would have been impossible to create in the real world safely. This meant the project could proceed to sea trials within months rather than years with the confidence and support of regulators and customers. The simulator also acts as a platform for operators and crew to get used to new interfaces, which is critical before sea trials.
Assist, not replace humans
Although an algorithm, AIM learns and replicates human-like decision-making. And humans can make mistakes. The first use case for deploying the system commercially is, therefore, not looking to replace the operator on board but act as an advanced assistant.
Hussain Quraishi, Manager, Strategic Innovation, Wärtsilä Voyage, believes there is a clear demand for AIM in some segments. “The value of AIM’s collision avoidance algorithms really comes into play in coastal shipping, inland waterways and ports,” he says. “We’ve seen a normalization of risk for some customers, for example, ferry operators in busy ports experiencing lots of near misses. They see this as normal, but it will eventually lead to loss.”
Quraishi expects AIM to be commercialized as a decision-support tool this year. This will mean that Wärtsilä Voyage will then be able to offer all three of the building blocks to full autonomy – decision making and logic, situational awareness, and action and control. However, he stresses though that full autonomy is not necessarily the goal.
“Our step-wise approach means customers may adopt the building blocks independently of each other, allowing them to start with whichever brings them the greatest real-world benefits first.”
To break it down even further, the scale of ‘manual’ to ‘autonomous’ vessel control can have different ranges, where system:
- Level 1 – Manually operated function, with no autonomous systems.
- Level 2 – decision supported function, where all actions are taken by human operator, but decision support tool provide options or otherwise influence the actions chosen.
- Level 3 – decision support function with conditional system execution capabilities, where some decisions and actions are taken by the system but need human supervision and acknowledgement before execution.
- Level 4 – self-controlled function, where system executes most of the operation with ‘human in the loop’, who can always override a decision or action.
- Level 5 – a fully autonomous function, which is completely unsupervised, and decisions are made and actioned by the system.
“And even within vessels, different systems can have different levels of autonomy. For instance, a system that alerts users about collision risks is level 2, whilst a system which autonomously steers the vessel to evade such risks would likely be at level 4 or higher,” explains Brightwell.
Right now, we are very far from level 5. However, there are many functions which are already on Level 3 and 4. Like AIM and those in IntelliTug.
The decision making and logic driving these systems are key components of Wärtsilä Voyage’s stepwise approach to vessel autonomy. And while these systems promise significant efficiencies for operators choosing to use it for decision support, there may be several steps yet before they can steer completely crewless vessels.
This post is sponsored by Wärtsilä Voyage. To know more about Wärtsilä Voyage’s stepwise approach to smart autonomy, download the white paper.