Latest / The Robotics Podcast with Fexingo: Autonomous Systems, Industrial Robots, and Hardware / Why Robot Arms Still Fumble a Slippery Fish
Transcript
- Lucas: So, Luna, I was watching a video this morning of a robot arm trying to pick up a raw salmon fillet, and it went exactly how you'd expect — the gripper closes, the fish squirts out, the gripper closes again, and the whole thing turns into a ten-second loop of failure. Luna: It's almost comical, but for anyone in food processing, that's not a joke — that's a real bottleneck. Lucas: Exactly. And it's not just fish. We've covered screwdrivers and doorknobs on this show, but today I want to drill into one of the hardest problems in robotic manipulation: gripping something that's both slippery and deformable. A raw chicken breast, a wet bar of soap, a just-caught fish — these objects fight back in ways a rigid block never does. Luna: And it's not just about friction, right? There's also the squish factor. Lucas: Right. You've got two failure modes. First, the surface is lubricated, so your gripper's normal force doesn't translate into shear force the way it does on a dry surface. Second, the object deforms under pressure, so the contact patch changes shape as you grip. The gripper's control loop is trying to hold a moving target. Luna: So what's the state of the art? Are there robots out there that can actually handle a slippery fish? Lucas: There are some promising approaches. One is to ditch the rigid jaws altogether and use soft, compliant grippers that conform to the object's shape. There's a company called Soft Robotics, based in Bedford, Massachusetts, that makes these pneumatic grippers with inflatable fingers — they look like little party balloons. They're used in poultry and produce lines to pick up items that would crush or slip. Luna: But I've seen those still struggle with really wet items, especially if the surface is super smooth. Lucas: That's the rub. Soft grippers help with deformation, but they don't automatically solve the lubrication problem. The contact area increases, which helps, but if the coefficient of friction is nearly zero, you still get slip. So researchers are turning to microstructures — patterns on the gripper surface that mimic nature. Luna: Oh, like tree frog toes. They've got those hexagonal channels that channel water away and create capillary adhesion. Lucas: Exactly. A team at MIT and Harvard published a study back in 2024 looking at exactly that. They fabricated these hexagonal micropatterns on a soft gripper and tested it on wet, slippery surfaces — including raw fish, I believe. The channels create a drainage path for the water, and the remaining thin film generates a suction-like effect. They reported a significant increase in grip stability compared to a flat gripper. Luna: So it's not just about friction — it's about managing the fluid at the interface. Lucas: Precisely. And the really clever part is the pattern is scalable and cheap to manufacture. You can mold it into existing gripper materials. That's a big deal for industrial adoption, because you don't need a completely new robot — you just swap out the fingertips. Luna: Before we go deeper, I want to give a quick shout-out to the people who keep this show going. We don't run ads, and that's because some of you have chosen to support us on buy me a coffee dot com slash fexingo. It's a small gesture, but it genuinely helps us keep the lights on and the mics on. Lucas: Yeah, honestly, if today's tech conversation gave you something usable, that's the link — buy me a coffee dot com slash fexingo. No pressure, ever. We're just glad you're here. Luna: And we're back to the fish. So, Lucas, the micropatterns are promising, but what does that mean for a real factory line that processes thousands of fish an hour? Lucas: That's where the rubber meets the road — or the silicone meets the salmon. The seafood industry is still heavily manual. Workers stand at conveyor belts, gutting and filleting, and it's tough, repetitive work. There's a shortage of labor, and automation has been slow to penetrate because the product is so variable. Each fish is a different size, shape, and slipperiness. Luna: So you need a system that can adapt in real time, not just a better gripper surface. Lucas: Right. The gripper is one piece. You also need sensors to detect the object's shape and slipperiness before you grip. Some labs are using vision systems combined with tactile sensors that measure local friction in real time. The robot can adjust its grip strategy based on whether the fish is wetter or drier. Luna: It's almost like the robot has to develop a 'feel' for the fish, similar to how a human would adjust their grip without thinking. Lucas: Exactly. And there's a company in Iceland called Valka that builds robotic systems for fish processing. They've been working on this for years. They use 3D vision to locate the fish and a specialized gripper that gently holds it without damaging the flesh. They claim their systems can handle up to sixty fish per minute — that's a one fish per second. Luna: That's impressive, but I wonder about the failure rate. If a robot drops a fish, what happens? Does it just get recirculated? Lucas: In most systems, yes, it'll loop back, but that costs time and can damage the product. So the industry is really focused on reliability. A 99 percent success rate sounds good, but if you're processing a thousand fish an hour, that's ten failures an hour, and each one might need a human to intervene. Luna: So the goal is to get as close to 100 percent as possible, which is a tall order with something as variable as a fish. Lucas: It is. And that's why researchers are also looking at active control strategies. Instead of a static grip, the robot might use a dynamic pinch — it closes quickly to make contact, then slightly releases and re-grips to establish a stable hold, much like how you'd pick up a wet soap bar. Luna: That's interesting — almost like a micro-vibration to break the surface tension. Lucas: Something like that. There's also research into using ultrasonic vibrations to create a thin layer of air between the gripper and the object, reducing adhesion — but that's the opposite problem. You don't want to repel the fish, you want to grab it. So the ultrasonic approach is more for handling delicate objects like semiconductor wafers. Luna: Right, different challenge. But for fish, it's about maximizing contact and managing the liquid. Lucas: And that's where the biomimetic microstructures really shine. The MIT and Harvard study showed that the hexagonal pattern not only channels water away but also traps a thin film that creates a suction effect. They tested it on a raw chicken breast and a wet glass plate, and the grip was significantly stronger than a flat control. Luna: I could see that being a game-changer for other industries too — like handling wet medical devices or slippery packaging. Lucas: Absolutely. The principles are universal. Any time you have a lubricated, non-rigid object, these techniques apply. And the beauty is, the pattern is easy to fabricate using standard molding techniques, so it's cost-effective for widespread adoption. Luna: So, what's the one thing you want our listeners to take away from this episode? Lucas: I think it's that the future of robotic gripping isn't just about stronger fingers — it's about smarter surfaces. By borrowing designs from nature and combining them with real-time sensing, we're finally cracking problems that seemed impossible just a few years ago. And the next time you see a robot fumble a fish, you'll know the physics behind it. Luna: And maybe one day, that robot will be the one filleting your dinner. Lucas: I'd be happy to let a robot do the slippery work. Thanks for listening, and we'll catch you next time.