Sparking a ROVolution

Elaine Maslin

October 1, 2015

Subsea robotics firm FMC Technologies Schilling Robotics has a vision, which it thinks could overturn the way subsea equipment is designed. Elaine Maslin found out more from co-founder Tyler Schilling.

Schilling Robotic’s HD camera. Photos from FMC Technologies Schilling Robotics.

Coming from the figure head of the foremost remote operated vehicle (ROV) firm in the oil and gas industry, the statement that today’s ROVs are clumsy is striking.

But, what’s more striking is the vision Tyler Schilling has for future ROVs. Today, FMC Technologies Schilling Robotics is working on HD cameras as sensors to enable positioning, as well as store embedded information, and shared control and tool dynamic positioning. But, in the future, Schilling has a more revolutionary goal – reversing the trend for subsea equipment to become ever more complex and designed to minimize ROV use. Having a new family of subsea robotics, not all free-flying, could enable subsea production systems to be more like simpler surface production systems.

There are a few more near-term goals he has first, however – including making ROVs, which are notoriously difficult to operate, less “preposterous” to pilot. “We still live in a phase of the industry that I refer to as the age of the exceptional individuals,” he says. “That means the largest determinant in the performance of one ROV crew over another is the guy behind the joy stick. It is ludicrously difficult to pilot an ROV.”

“The thing that makes it even more preposterous, even in today’s depressed environment, is that there are operations out there costing more than $600,000/day being held up by people fiddling around with a machine like this. I think they can be made 4-6 times more productive then they are now.”

Tyler Schilling in the Schilling Robotics assembly shop, Davis, California.

Schilling Robotics is working on a number of areas to make that happen. “One of the main technology threads we are pursuing is one we call shared control,” Schilling says. This is a term long-used in the robotics industry to denote technology where you share duties between the machine and the human, playing to each’s strengths. This has been illustrated by so-called “peg in hole” tasks – where a robot needs to align an object to a hole and mate it. To perform such a task with an ROV, some 10 degrees of freedom need to be aligned – which is why ROV work is so hard. If a computer was used to perform the alignment, with the human supplying the intent, the operation would be easier.

This is what Schilling Robotics calls tool dynamic positioning. “We are just about ready to try our Beta release of tool dynamic positioning to the market place and do some offshore tasks with it,” Schilling says. “We realized it would be profoundly beneficial if, rather than the body of the ROV being positioned relative to the sea floor, the ROV could be positioned relative to the well or the work space. That’s really what the user wants. All we have done is make his or her job incrementally easier by positioning the body of the ROV. That’s a big area we have been working on.”

Another area Schilling thinks great strides can be made in the sector is by using video cameras as sensors, “going beyond just letting people look at stuff, but using it as a sensor to control the motion of both the ROV and its tool,” he says, turning the camera into a multi-functional instrument.

What he is describing is the technology that allows car drivers to sit back while their car parks itself, or skiers to set off a drone, which follows and films them as they go down the mountain side. It is charge coupled device (CCD) technology, which uses an integrated circuit etched onto a silicon surface, forming light sensitive elements, which record light by electronically reading the charge created by photons which land on its surface, turning this into a digital image – and data. CCD has given rise to SLAM, or simultaneous localization and mapping, which, using fast processing algorithms can extract fiducial points from each frame of a video and compute the motion of those fiducials from frame to frame, allowing the user to create a 3D map of the object and plot the position of the camera relative to it at the same time.

“We think we are at the very beginning of a large advance for the industry,” Schilling says. “Bring this technology together with the ROV and its tool and dynamic positioning and you have a machine able to continuously understand where it is and have an ever improving 3D model of the work space. This will see productivity and ease of use take off.”

The realization around the role video cameras can play in improving ROV productivity led Schilling Robotics to invest in video camera technology. The firm’s camera sends data back along the same Ethernet as the data and communications for the ROV uses, filling the video frames with meta-data, which can then be searched, such as depth, heading, etc. This information, instead of just being live on a screen, is now machine readable and searchable.


Schilling says the firm is also considering the implications of 3D printing and the potential to design in the option to have parts both able to be mass produced and 3D printed on site. “We have been discussing why don’t we build in to the development process a rule that says think of ways to design parts so they can both be produced by 3D printing and normal manufacturing processes, so we can make parts economically but they can also be printed on the spot,” Schilling says.

But, by far the biggest innovation is yet to come.

“One of the most exciting threads we are focused on is actually a long-term horizon one,” Schilling says. “Right now we are investing in shared control with video as a sensor and a whole host of technology in that vein. But, it occurred to us a couple of years ago that the people who design equipment that produces oil on the sea floor have typically been designing equipment that uses the ROV as little as possible.

“As an example for flowline connectors, the task has been reduced to simply require the ROV to pump fluid into the connector or the tool in order for it to latch. This certainly asks for very little from the ROV; however, the ROV intervention makes for a large and expensive connector. We believe and have started a program to turn that model on its head. We’re asking what new family of robots would enable sea floor equipment designers to make substantially less expensive equipment.”

“The poster child in the surface industry is to hook up a flowline using an API flange. There’s a flange on both sides, a gasket then a plurality of nuts and bolts. It is less than 1% the cost of an ROV mateable sea floor connector, largely because the robot cannot deal with all those parts. That, in my opinion, is a challenge.

“I think it will take five years for the first ‘thing’ to hit the water and maybe 10 years for it to really pick up and take off. I visualize it as a whole new family of robotic devices, of which only a subset would be free-flying ROVs. It is very conceptual at the moment and what we are trying to do right now is envision what the fundamental principles would be that would drive the eventual design.”

But, will the industry be receptive? “I have spent the last 30 years with people telling me ‘Tyler, that stuff is too fancy, all we need is agricultural technology, and you keep pushing stuff on us that’s too fancy.’ But I’m pleased to report, as of two years ago, I now agree with those people because the agricultural industry has surpassed us in the use of ‘fancy technology’. The level of automation in the agriculture industry is remarkable.”