Imagining the next generation of robofish

Follow Our Changing World on Apple Podcasts, Spotify, iHeartRADIO or wherever you listen to your podcasts 

They will look like fish, swim like fish and even sense like fish. In future we might have robofish patrolling our reefs and protecting our harbours. 

Robofish for ocean monitoring 

Professor Iain Anderson leads the biomimetics lab at the Auckland Bioengineering Institute (ABI), where many different projects looking at improving our ability to monitor our oceans are underway.  

“It would be fair to say we know more about the surface of Mars than we do about the surface of our own planet because you can't see through the ocean,” says Iain. “You've got to go down there and look. You’ve got to go down there with something to see.” 

It’s an area of science and technology that Iain is passionate about. 

“We need to be able to know what the status of our reefs are,” he says. “We're getting invasive species coming in [but] we don't know they're here until it's too late. We just don't have a good hand on the pulse of what is happening in our own aquatic environment.” 

Taniwha 

One of the ABI’s early forays into machines that can explore the ocean floor was a human-powered submarine called Taniwha. While it won some prizes at the international submarine races in the US nearly a decade ago, the work happening in the lab today is aiming to take the human out and put the robot in. 

Iain says the idea is to go from robots which are steel and hard plastic to robots made from silicone and soft materials. But also, to give them more lifelike attributes – such as the ability to sense the environment around them, and quiet actuators that move and create force without gearboxes.  

“Fins are quiet, so they don't disturb marine life, and they also don't get clogged up with the weed,” said Iain. “If you have a little propeller, it sucks the weed in and it gets jammed. You're not going to get your fingers chopped off in a fin.” 

Moving like a fish 

It is this propulsion part of the robofish that PhD candidate Robin Milward Cooney is working on. He is developing artificial muscles, which are essentially two layers of plastic with liquid inside.  

“This forms a capacitor and the fluid inside gets pushed to one side of the membrane and then depending on the shape of the polymer sheets you can make different motion by moving the fluid around the different channels,” says Robin. 

His work, funded by the Ministry of Business, Innovation and Employment (MBIE), is to make a quiet submersible that can be used to patrol the reefs of New Zealand, without scaring the fish away. 

Sensing like a fish 

It’s one thing to move like a fish, but Arne Bruns, another PhD candidate, has funding from the Office of Naval Research to get robots to sense like fish too. 

“If you want to swim like a fish, you have got to sense like a fish and that’s a big challenge,” he says. 

To tackle this problem, Arne has had a custom-built sensory testing tank set up to help him develop his sensor. A propellor sends water around the tank, at speeds up to about two metres a second, past his fin-shaped sensor which provides measurements back to his laptop.  

“So we can test something, correct something in the sensor, come back, do it again,” says Arne. 

The sensor itself looks like a fish tailfin but made of clear plastic with a black strip around its edges. 

“It's essentially just silicone and carbon,” he says. “And then extreme precision in manufacturing to be able to get the layers so thin.” 

Arne makes inks from conductive carbon and then blade casts them to 50 microns –which is about the width of a single strand of hair – and the ink creates a capacitor inside the silicone fin. 

“So, the membrane itself is the sensor. It's like three electrodes stacked and as this membrane stretches from the water flow, the capacitance, the electrical properties, change and we can pick that up,” says Arne. 

Iain says fish fins have rays that they can actively spread to change the tension, and Arne's fin is trying to mimic that. 

He is also investigating how that change in tension and the number of rays influences the sensing ability and whether the thickness of the membrane has an influence too.  

“Basically investigating how can we best design these fins to give us the sensation that the fish have in their fins,” he says. “[The] goal is that if a fish propels itself with its fins, we want to know how much force did it put on the water, how much are we moving forward, how effective was this, how much energy have we used?” 

And he says they are learning this through the raw data collected by the fin.  

“So we know the water speed, we know the area, we can do fluid simulations of how much force is being exerted onto the membrane and then we can look at our data and say, okay, if we get this signal, that means this amount of force onto the fish, then if we install the sensor in the fish, we get a signal that we know what it means in terms of force.” 

Why do we need robofish that can sense? 

According to Arne, the commercial robofishes currently on the market – for show or for film – only have about an hour battery time. But he said with his sensors, you can get feedback and adapt swimming in a way that saves you a lot of energy. 

“If you want to go deep or you want to go far or want to spend time underwater you need high efficiency,” he says. 

Arne can envisage putting a fish robot complete with sensors in different scenarios and using machine learning to teach it how to react to outside input.  

“The goal of my PhD is basically to build a cornucopia, a big bread basket of sensors and put as many different sensors on the fish to get this very chaotic data from all the different sensors and when it recognises the same kind of pattern again from all the different sensors then we can programme it how to react to this condition.” 

Arne says there are multiple applications for robofish, from surveying reefs, to checking underwater infrastructure and even to patrolling our shorelines for security breaches. 

“You need things that can swim up and down, long distances. [They] need to be very energy efficient and having silent observers that can alert you to potentially threatening situations would be quite helpful,” he says. 

A smart wetsuit 

Building sensors and muscles for robofish are not the only projects happening at the biomimetic lab. Dr Derek Antillon Orbaugh and Cheng-Huan Lu have been working on making smart diving gloves and wetsuits. 

“The idea is to enhance diver safety during a mission for that we're developing this vest that will monitor the physiological parameters of a diver,” says Derek. 

Once fully developed, the suit could provide feedback to ensure the diver is swimming efficiently or tell a dive buddy if they are in trouble, through sensors that monitor swimming style and specific hydrogels that can measure heart rate underwater. 

Find our more about the wetsuit and the robofish by listening to the episode. 

Learn more: 

• Claire Concannon checked out more of the Auckland Bioengineering Institute’s research last year in Digital twins and beating hearts. 

• The podcast Voice of Tangaroa takes a deep dive into the state of New Zealand’s oceans.