No GPS Underwater: How the Pool Robot Learns to Listen

At the bottom of a suburban pool, a box the size of a satchel glides along the tiles, climbs a wall, scrubs the waterline, then drops back down. It moves without a cord, without a pole, without anyone at the edge. And it moves blind: a few centimeters below the surface, the signal that guides a car, a drone or a robot mower has already vanished.

Water stops radio waves. GPS, which places an autonomous mower to the centimeter in a garden, does not cross the surface of a pool. The robotic pool cleaners shown in 2026 therefore had to solve a problem their land-bound cousins are spared: finding their way inside a closed volume, with no sky and no fixed landmark. Their answer comes in a word borrowed from submarines, sonar.

Why Positioning Stops at the Surface

The first pool robots simply bounced from one wall to the next, like a puck, betting that enough zigzags would eventually cover everything. The method worked on a small rectangular pool, far less on a kidney-shaped one with steps and a deep end. To clean quickly without missing a corner, the machine had to know where it was.

Out of the water, the satellite settled that question long ago. Underwater, everything changes. Radio waves die within centimeters, so GPS is useless. The gyroscope, which measures rotation, drifts and piles up error minute by minute. And an optical camera on its own runs into glare, murky water and the distortion that surface ripples cast on light.

That leaves sound. An ultrasonic pulse travels well through water, bounces off the walls and returns; by timing the echo, the robot works out its distance to each wall to within millimeters. This is sonar, long reserved for submarines, which makers such as Seauto were the first to bring into the backyard pool. Paired with an inertial unit and, now, a camera, it lets the machine map the floor and plan a route, much as a robot vacuum maps a living room.

That intelligence stays on board. The calculation cannot be handed to a remote server: underwater, the wireless link does not pass, and obstacle avoidance plays out in a fraction of a second. So the robot decides alone, on board, which puts it closer to a miniature self-driving vehicle than to a remote-controlled vacuum.

A Summer Chore That Drops Off the Calendar

Maintaining a pool by hand is a weekly ritual: skim the surface, brush the walls, hook a suction pole to the outlet and walk it across the bottom for half an hour, start again after every storm. The 2026 robot absorbs all of it. It handles the floor, the walls and the waterline, schedules itself from an app and returns to its dock to recharge.

The numbers show how far the technology has come. The Aiper Scuba V3 claims up to three hours of battery life, suction on the order of 4,800 gallons per hour and filtration that captures particles down to three microns, well below a grain of pollen. The WYBOT C2 Vision pushes runtime to five hours; the MOVA Diver A10 boasts 6,000 gallons per hour across three brushless motors. All belong to the cordless class, free of the cable that used to tangle on the bottom.

• Aiper Scuba V3: underwater camera and optical sensors, up to three hours of runtime, contactless charging on its dock.
• WYBOT C2 Vision: onboard camera to spot debris, up to five hours of operation, sonar obstacle avoidance.
• MOVA Diver A10: triple brushless motor, 6,000 gallons per hour, dual-pass waterline scrubbing.

The market reflects the appetite. Valued at around 1.1 billion dollars in 2025, it should top 1.25 billion in 2026, carried by annual growth close to 14 percent. What these devices sell is not the hydraulic feat, it is the Saturday morning handed back: clear water with no gesture and no decision, upkeep that happens while you sleep or work.

A Camera Filming Underwater, in the Backyard

To find their way and tell a dead leaf from an insect, the high-end models have gained eyes. The Aiper Scuba V3's camera sees up to two meters and has been trained to recognize more than twenty kinds of debris, so it can come back and work the spots where dirt gathers. The precision is real. It carries a familiar flip side.

That camera does not stop at the water's surface. It is tied to an app, often to an online account, and it sits, by definition, in a private space: the garden, the terrace, the back of a house. Manufacturers say the images serve only navigation and stay on board. The promise is likely kept. But it rests on their architecture, not on any control the owner exercises, and you have installed one more connected object watching an intimate corner of the home.

The Price of Comfort, and a Recall to Underline It

The first hurdle is the price. A model with vision and sonar runs between 900 and 1,200 dollars, several times the cost of an ordinary suction cleaner. At that price you are not buying an accessory but a device the pool's upkeep depends on, with its updates, its app and, sometimes, features locked behind a subscription.

Then comes the battery, the cost of going cordless. In March 2025, the U.S. product safety regulator recalled an Aiper model, the Seagull Pro: more than thirty-five thousand units involved, nineteen reports of devices melting, smoking or catching fire while charging, five cases of property damage. No injuries, but the recall underlines the obvious: a large lithium battery that heats up beside the water is no trivial object.

Then there is the ordinary dependency. The map of the pool, the schedules, the history live in an online account. Smooth running assumes a maker that keeps its app alive, does not change its terms and does not disappear. You delegate a chore; you inherit a relationship, made of trust and possible failures, with a company and its cloud.

Water That Stays Clear, Within Limits

The 2026 pool robot keeps a concrete promise. It erases a seasonal chore, sets up in minutes and keeps the water cleaner than irregular upkeep would. For anyone with a pool and little time, the call is easy to make.

What you gain is measured in afternoons: the ones you no longer spend, net in hand, chasing leaves. What you give up is more diffuse. A budget first, a battery not to be ignored, and the quiet habit that settles in, of entrusting a corner of daily life to a machine and to the company steering it from afar. Sonar taught the robot to find its way where no satellite follows. The question left is whether we keep an eye on what the machine, for its part, sees so well.