Humans have made their mark all over the world, sometimes with disastrous consequences, although we’ve still only explored about a third of the planet. More than 70 per cent of the Earth’s surface is covered by oceans and only 5 per cent of the watery mass has been explored. The challenge is more than just how much of the underwater world we might map – it’s also about how we can do a better job looking after the things we put there.In 2014, Nasir Ahsan began pondering a bigger oceanic future while aboard a scientific research vessel cruising the Caribbean and the Atlantic. The mission at the time was to search for hydrothermal vents and mud volcanos, although it would also inspire a new kind of business for the young man working towards a PhD in Marine Robotics and Machine Learning.Nasir says that as the navigator of a Remotely Operated underwater Vehicle (ROV) he was “looking at things no one had seen before. We were pushing limits and I was in the front seat… but underwater the biggest problem is you can’t really see very far ahead. They use sonar, but lines attenuate [degrade] and signals attenuate, which is why you don’t have communications.”When Nasir returned to shore, he was all fired up about breaking through the four barriers to science and work underwater: Power, communications, autonomy and perception.
Can machines handle the work alone?
It came to Nasir that one possible way of improving how submerged machines operate was to apply machine learning to the problems he’d encountered.
“What if an autonomous robot, or a fleet of robots, could work independently using algorithms and sensor data?” he thought. “One of the biggest challenges is that you can’t position under water without using extremely expensive acoustic positional systems.”
Autonomous robots may sound like something out of a classic Isaac Asimov sci-fi novel, but the costs of conventional techniques used for offshore underwater work are equally out-of-this-world.
Nasir is something of an evangelist for the new technology coming online that is revolutionising underwater positioning – and saving organisations that work undersea a fortune:
“Big ships cost about $120k a day to power with diesel fuel, so a week offshore for a ship and a crew is in the millions. The move towards autonomy can be a big time-saver that can allow crews to focus on other things ‑ or not even be on the ship. Autonomous robots could do the work!”
Out of the deep is born Abyss Solutions
Nasir’s underwater eyes are being refined by his university friend and optics expert Masood Naqshbandi, who was working on his Masters in Materials Chemistry and Photonics.
“I think when we met at uni, Nas was finishing his post doc…” begins Masood.
“More like getting redundant!” interjects Nasir.
“Anyway, he says to me: ‘Dude, I think the hardware is done’,” remembers Masood. “And I’d been building robots since I was a kid in Afghanistan ‑ I had this old toy building set from the Soviet Union. It wasn’t nuclear-powered although it wasn’t safe or child-friendly. It had everything like nuts and bolts, wheels, cogs, metallic parts (possibly too sharp), the whole lot for building different things. So I made robots.
”Nasir jumps in to explain that Masood has much better hardware to work with now they’re running Abyss Solutions. “The robot is just a mobile vehicle for the sensors,” he says. “How you make sense out of those sensors and how you react to that data, that’s where the next step is.”
Their first task is to tackle the vital issue of improving perception underwater using optical sensors. It’s not so much about building robots at Abyss Solutions, it’s more about decking them out with an increasingly elaborate array of sensors and helping customers use that data.
“The sensors are really important because they help us get the right photogrammetry data for contrast and resolution,” explains Masood. “We can then take it into 3D modelling software, correct for the distortion and begin interpreting what’s in the environment.”
Getting in and out of hot water
Nasir and Masood’s first robot project literally landed them in hot water. They built a pen-sized robot with a simple camera that autonomously inspects hot water systems.
“We were business novices,” laughs Masood. “We followed this one client who said, ‘Wouldn’t it be great if you could put robots in the hot water systems?’ They have a solution to a problem no-one cared about”.
Fortunately, the autonomous robot inspection pitch did help Abyss Solutions gain traction with clients who had far bigger bodies of water to deal with and far bigger problems. Business began to look good when Abyss landed a job with Sydney Water to inspect Alexandra Canal.
Nasir explains how the canal inspection job was the market validation they were after. “Those are particular canals, they are heritage-listed, 150 years old, you can’t fire your drones over there because of airspace restrictions as it’s right next to the airport. Industrial waste is dumped in the water so it’s not a safe area.
”The success of that project led to more work with Sydney Water and the team is delighted to report it has one of the biggest water utilities in the world on board ‑ with the potential for repeat business.
What treasures await?
“The kind of customers we have are people with assets that have problems or are ageing ‑ like dams, waterways, canals, ports, marinas, marine wharves, jetties, bridges, submerged pipelines,” says Masood.
“The only solution they could find was using our robots,” adds Nasir. “Our data can help save a lot of money when they’re specifying maintenance tenders because it’s very accurate, targeted and precise.”The duo acknowledges that being able to firm up longer-term relationships with early clients gives them the assurance to rethink the big picture:
Masood believes techniques and technologies they’re working on now will scale well beyond capturing data on structures like sea walls or jetties. “Our bigger vision is to take this technology not only to limited infrastructure and limited assets, but to apply it generally to all the underwater area.”
“At the moment we’re inspecting things,” agrees Nasir. “Then… with this technology, island-building, harvesting energy, transporting that energy to other locations all become possible. Building surveillance hubs underwater becomes possible ‑ marine hubs where scientists can study, mining companies can explore, search for minerals, base stations can be established…”
The question now is what’s stopping scores of businesses from rolling out massive underwater infrastructure projects?
It comes back to the optics, says the pair. In the next two to three years, they expect to have access to super cameras capable of imaging 20-25 MTU scores (MTU is the measure of turbidity, murkiness of water). Organisations with underwater aspirations are looking for optics with better MTU scores because the better they can see underwater, the more they can do.
“The priority is to understand where the product fits and what the market wants before we skill, so we’re looking at multiple sectors,” says Nasir.
“The challenges of getting good data and images aren’t limited to underwater – they’re also challenges in space and on land,” says Masood. “Autonomous cars, for example, don’t use visual cameras for navigation because they can’t see through turbulent conditions (rain, fog, smog) so instead use acoustics and sonars. If we can solve that underwater, we solve it on land and even in space.”
Just don’t expect to find a robot in your water heater any time soon.
muru-D helps build businesses
Abyss Solutions came to Telstra’s incubator muru-D with a fair amount of hothousing experience as alumni of Sydney University’s Incubate and Australian Technology Park (ATP). As Nasir recalls, it was clear from the outset on why they wanted to go with muru-D.
“One of the questions Mick (Liubinskas, a muru-D mentor who grew Pollenizer and Kazaa) asked me was, ‘Why do you need us? You’ve got clients, you’ve got traction, and you know what you’re doing’. I said ‘Help our business get on the road, we don’t know how to run a business, we’ve just got technology’. He helped us become a business.”
Here are the highlights of what Abyss learned at muru-D:
- muru-D helps startups understand their customer requirements – “It’s something muru-D really helped us with,” says Nasir. “Getting on the customer side, not being shy, talking to people, getting feedback, growing. These were some of the basic skills we learned that really helped the business. Now we actually know it’s not just the technical value proposition; we know what a commercial value proposition looks like and we can frame our business in that context and grow.”
- muru-D offers a strong, valuable business network – “You end up having the right people around the business to guide you,” says Masood. “The muru-D and Telstra brands open a lot of doors as well, plus channel partners like 7West Media.”
- muru-D’s experts get the day-to-day challenges –“Ian Davis really helped us a lot sorting out corporate governance and with a high level of business strategy, ” recalls Masood.
Abyss solutions’ top tips for startups
- Sell, sell, sell – “You sell, you’re good, you can deal, you can bargain,” says Nasir. “If you don’t sell, you’ve got no bargaining power and obviously you also don’t understand your customers, you don’t understand your product. You don’t know what you’re building, you’re still stuck in the tech world, you need to get out of it. If you’ve got a very generic technology, you’ve got to make sure you find the right vertical before you scale, make sure you understand all these things. So sell, sell, sell ‑ that is extremely important. Understand the different kinds of clients, talk to different markets.”
- Think for yourself – “Have an open mind when it comes to ideas, but don’t absorb everything,” warns Masood. “Be strong in what you believe and what you want to follow, but base that on the advice you get. Be critical when you get advice, because you can’t follow a million different paths at the same time.”
- Don’t swim with sharks… but don’t hide away, either – “There’s a word in business: ‘Sharks’. Stay away from those,” says Nasir. “Stay away from bad deals and from bad channel partners. How much commission do you give to a channel partner? It needs to work for you.”“Stealth mode doesn’t work anymore,” adds Masood. “You can’t just code at home, for months on end thinking you’ll come up with an amazing product. You need to get out there.”