In the 1980’s, remotely operated vehicles (ROVs) were revolutionised with the introduction of fibre optic cables to replace the old and outmoded copper ones. This allowed a far greater quantity of data and video content to be transmitted from the ROV back to the surface. In addition, it paved the way for ROV’s to function at greater depths than was ever previously possible.
Since then there have been plenty of incremental advances and iterative improvements in the ROV sector but there hasn’t been one single identifiable revolutionary shift, until now.
Thanks to a perfect combination of advances in automation control, machine vision technology and the development of new algorithms for artificial intelligence, the whole industry is on the precipice of a major step-change.
At this stage it would be virtually impossible to create a fully automated ROV, that was guided by artificial intelligence that could pre-empt and react to changes in the underwater environment without human interaction.
However, whilst a fully automated, reactive and self-aware ROV is not feasible, it is possible to automate many of the day to day ROV tasks to reduce the strain on the pilot.
By breaking down a routine operation that an ROV may perform into little steps, and automating each of these steps, you can gradually build up an ROV’s automatic capability. the pilot can simply set the ROV to this task, instead of having to manually pilot the ROV the entire time.
Whilst the ROV is not completely self-aware, the automation program can tell when one of the steps it has to perform to complete a task has been interrupted, by unforeseeable circumstance, and can immediately alert the human pilot who can re-assert control of the ROV and guide it out of trouble.
By allowing the pilot to simply oversee and manage the ROV as it completes quotidian tasks by itself, it reduces the stress of the pilot and allows them a greater degree of focus for more critical tasks, thus increasing safety and reducing the likelihood of incidents.
Some of the most difficult ROV tasks require extreme precision, manoeuvring the ROV to a very specific distance away from an object, and maintaining that distance. These delicate tasks can be made more straightforward when the ROV is fitted with an automated navigation and control system. This allows the ROV to maintain its position in the water by ‘seeing’ where it is in relation to other objects, enabling it to remain in a stable position, even in turbulent offshore environments.
These types of automated navigation systems utilise a ground-breaking combination of both Vision Technology and ROV motion dynamics. Whilst Vision Technology uses a camera to track the relative position of the ROV in relation to subsea objects, ROV motion dynamics uses a wide variety of motion sensors that allow the ROV to understand and anticipate how it needs to move. Applying the brakes when it approaches a target position for example.
By setting an ROV to automatically maintain a certain distance from an object, the ROV pilot can focus on the task at hand, rather than have to divert half their attention on keeping the ROV stable.
In a typical ROV manipulator, there are six degrees of movement that can be achieved. The three cartesian motions (the x, y and z axis) and the three angular motions (pitch, roll and yaw). Whilst it can be advantageous to be able to control all six of these degrees of motion at the same time, it is not always desirable.
Manipulator control methods rely extremely heavily on the skill of the operator, therefore a system where some of the process can be automated reduces the level of operational stress on the ROV pilot. This in turn contributes to a safer and less incident prone work site.
Recent developments in the sector allow for exactly that, ROV operators can now select which of the six degrees of movement he/she want to control, and let the software handle the others. This means that the human pilot can still control the most critical movements specific to that job, whilst the software can ensure that the less important motions are taken care of, thus increasing productivity and efficiency.
Up until now the foundation of fault detection in ROVs have been based upon sensors that test for low, high or band range limits. Whilst the advent of digital control systems has allowed for these sensors and their limits to be adjusted in real time, the system is still too rigid and inflexible to keep up with the spiralling level of complexity associated with modern ROVs.
With each additional function that is added to an ROV, more of these sensors must be added to monitor each of these functions. Based on this it can easily be seen how a model of a single sensor to detect a single fault can quickly become unworkable.
A better alternative to the conventional sensor system is an Intelligent Diagnostic System. These systems use basic equations of electrical power flow through the ROV, and combine it with telemetry from sensors to create a virtual model of the system. It can then use this to more accurately diagnose faults in an ROV, by pulling all the salient information into one central place.
Using a system like this, multiple simultaneous failures can be more easily distinguished from each other and the root cause, that may have caused multiple failures can be identified much more easily.
With the introduction of low-earth to orbit satellites, offshore sites can now access high-bandwidth, low-latency networks, worldwide. Combining this with the latest computer programs that aid with the control of an ROV, means that ROV’s can be operated remotely, in real time, with the pilot not even required to be in the same country as the ROV.
This creates obvious cost saving opportunities, due to the fact that the pilots no longer have to be in close proximity to the ROV that they are controlling. It also allows for experts in specific fields of ROV piloting to be rapidly deployed during incidents or in times where high skill are critical to operational needs.
This development means that ROV’s are one step closer to being truly remote operated, with the range of the pilot extending from a nearby vessel, to anywhere with a secure internet connection.
Looking to the future
As this article covers, there is tremendous potential for innovation by the incorporation of AI technology into the world of ROVs.
Moreover, all of these separate innovations are starting to converge and mesh together into something greater, an artificial intelligence revolution.
In the future there is further scope to create fully autonomous ROV control programs that can pilot themselves to underwater coordinates, follow pipelines with machine vision, and even avoid objects on their route.