From Shadows to Certainty

They were not supposed to be doing this. Swarnab Banerjee is a power electronics and motors expert. His younger brother Rishav is a mechanical engineer from automotive manufacturing. When the two of them founded Banergy in 2021, the plan was straightforward: create a reliable channel to get quality marine robotics components into the hands of Indian engineers and researchers who had no good way to access them. The MATE ROV competition had shown Rishav something troubling: Southeast Asian and Indian student teams were technically capable but consistently disadvantaged. The components available to them were inferior, costs were prohibitive, and there was no local knowledge infrastructure to help them use what little they had. Banergy would fix that. Supply chain, support, hand-holding. That was the mission.

Then 2024 happened.

Rivers overflowed. Bridges collapsed. Buses fell into floodwater. Boats capsized on rivers swollen with monsoon runoff and unannounced dam releases. The scale of water disasters overwhelmed emergency services still relying entirely on manual methods. And Banergy — a company that had set out to sell components — found itself on the phone with Indian government disaster management officials who had a single request.

We need a camera underwater.

The First Lesson: Simple Problems Are Never Simple

Swarnab thought he understood the requirement. A waterproof housing, an action camera, a tether. He went to the first field trial confident they had it covered.

The river changed his mind within minutes.

The Ganga — Banergy’s home territory in the Bengal delta — is not a forgiving environment. Optical visibility runs to ten or fifteen centimeters. Tidal bores push waves upstream as the tide rises. Dams release water without warning. The river carries heavy silt, debris, and the residue of millions of people who bathe and cremate along its banks. Action cameras, underwater cameras, any optical system — all effectively blind.

Cameras don’t work here. Optics will not work. Sonar was our bet.”

- Swarnab Banerjee

But sonar brought its own challenges. In the open ocean, turbulence fades below a certain depth and a sonar unit stabilizes. In a tidal river during a flood, turbulence runs constant through the entire water column. Signal degrades. Sediment creates noise. And if the vehicle can’t hold position against the current, the imaging becomes unusable regardless of the hardware’s capability. Solving for sonar meant solving for everything around it.

Building the Answer from the Ground Up

Banergy’s starting point was a modified ROV, re-engineered for stability in turbulent Indian river conditions — because even the best sonar is useless on a platform fighting the current.

For the sonar itself, cost was not secondary — it was central. India’s disaster management sector operates on budgets where high-end multi-beam systems simply cannot scale. Traditional high-end equipment that dominated industry conversation delivered the performance but the economics did not work.

Cerulean changed that. Swarnab had first encountered the forward-looking sonar through a video on Instagram so ordered a unit to see the results firsthand before any operational mission depended on it. That firsthand knowledge was the confidence. Whether it would hold up in a silt-heavy, turbulent river rather than the open Pacific — that was still a leap. But looking at the science, and looking at the absence of any better affordable option, Banergy made the call.

We are not going to be able to control the environment, so we might as well take a bet on the technology.”

- Swarnab Banerjee

Above the waterline, the brothers faced a different problem. A laptop in the field — in rain, in a thunderstorm, at 4am on the bank of a flooding river — is not a command platform, so they built one. The Surface Command System is an edge-computing unit pulling sonar data, camera feeds, vehicle telemetry, and IMU readings into a unified display. Operators interact with it like a video game: flight stick, port controls, a Meta Quest VR headset recording what the operator sees alongside the ROV feed for mission debriefs.

Layered across all of this runs Centurion, Banergy’s proprietary AI. Rather than building a separate interface and fighting to integrate competing data streams, Centurion watches Cerulean’s SonarView software display — exactly as a human operator would — and runs object detection continuously in the background. The human watches one thing at a time, the AI watches everything at once. By overlaying their capabilities on SonarView rather than rebuilding from scratch, Banergy compressed what might have taken years into months.

What Results Actually Look Like

Before Banergy, a search-and-recovery operation in a tidal Indian river meant identifying a witness, making a best guess at the target location, and sending divers in on an expanding radial search pattern — repeated within the narrow tidal windows, a few hours either side of slack water, when diving was survivable. At night, everything stopped. A diver entering zero-visibility water with a fifty-fifty chance of finding anything was simultaneously risking their life and watching team morale erode. Typical mission time: three to four days.

With the system in place, the longest single mission Banergy has run was reduced to hours. Operations that previously consumed days have been completed in under two. Robots work at night. Robots work between tides. Human risk during the search phase approaches zero.

In Operation Basanti — recovery of a 30-foot trawler capsized in a thunderstorm — the system did not simply locate the vessel. It delivered its GPS coordinates, its orientation on the riverbed, which side it had tilted onto, and exactly where divers should attach recovery lines. A diver descending without that information is searching. A diver descending with it is executing a plan.

For a diver going into zero visibility, this changes everything. Instead of searching blindly, they’re executing a precise plan.”

- Rishav Banerjee

Another striking demonstration came not from a vessel recovery but from a search for lost equipment. A high-value sonar unit had been lost in the river. The disaster management commander estimated 72 hours to recover it. Banergy deployed an ROV with a Cerulean sonar unit, located the device, identified its condition, and traced its 20-metre cable drifting in the current. The unit was back in people’s hands in four and a half hours. The operation had started at 4am in a thunderstorm.

Sonar rescued sonar. The industry noticed.

The Problem Nobody Was Solving

After the hardware proved itself, the next barrier emerged. Operators unfamiliar with acoustic imaging looked at the screen and saw shadows. They asked why the boat didn’t look like a boat. Without proper interpretation, even the best system risks being dismissed at the moment it is needed most.

This, Swarnab argues, is the defining failure of the marine robotics industry: everyone sells products, almost nobody sells the solution that produces a measurable result.

Banergy built that knowledge infrastructure themselves. The training programme — Made Under Pressure — takes operators from ROV and USV fundamentals through sonar interpretation, search strategy, and field simulation. Underpinning it is RoTOPS: Robot Training and Operating Procedures Standardisation, modelled deliberately on NATOPS, the American Naval Aviation standard developed for F-18 operations. When two certified operators from different teams meet at an incident, they don’t negotiate procedure. They share a standard, share a language, and get to work. Training cohorts run at 25 to 50 people — the scale at which Indian government programs operate.

Everybody’s been selling product, but nobody’s selling the knowledge of how to use that product to an economic benefit.”

- Swarnab Banerjee

Three Layers, Built to Scale

The operational model is a layered system. Layer one: Banergy unmanned surface vessels deployed immediately on arrival, running Cerulean side-scan sonar to map the riverbed and flag areas of interest while the rest of the team is still setting up. Layer two: the Banergy ROV, investigating specific targets with Cerulean forward-looking sonar, pinning precise coordinates and building a complete picture of what is on the riverbed and in what condition. Layer three: the human diver, equipped with a diver propulsion vehicle; Banergy DPV, descending with a plan rather than a guess. Integrating the three layers, on top Banergy Centurion AI assisting operations.

Each layer reduces the risk and increases the efficiency of the next. Cerulean’s sonar threads through all three: compact enough to mount on Banergy USV, integrated into the Banergy ROV, and priced at the point where deploying at Indian government scale is actually viable.

A Market Being Created

Banergy’s longer-term vision is preemptive. India’s sacred river ghats — the stone bathing platforms along the Ganga — are existing infrastructure at exactly the locations where water incidents occur. The goal is to have robotic systems permanently stationed there, autonomous and ready, already in the water before the question of jurisdiction has been answered. That is how lives get saved: not by faster human response, but by eliminating the delay between incident and action entirely.

None of this was the plan. They came from electric vehicles. Disaster management arrived uninvited, and they approached it the only way they know — straight to the source, fix the foundation, not the surface.

This is a market Banergy is building, not entering. It requires partners willing to go on the journey — companies like Cerulean, whose technology makes the economics viable and whose capability in the hardest environments makes the results real. Indian disaster response forces already deploy across Bhutan, Nepal, Sri Lanka, Thailand, and the broader subcontinent. Banergy’s solution is built to travel further.

The families on the bank are still waiting. But the system reaching them is already proven, already scaling, and moving toward the moment when it arrives before anyone has had to be told to send it.

Learn more from Cerulean Sonar