ISRO Taps AI to Spot Deadly Rip Currents Off Andhra’s Coast

The Space Applications Centre of ISRO has joined hands with Andhra University to build an AI-powered system that detects rip currents on Visakhapatnam’s beaches before swimmers wander into them. The two-year effort, called Project Bharati, runs on a budget of Rs 22 lakh and pairs a satellite-to-shoreline data pipeline with 360-degree AI cameras mounted on solar-powered mobile units.

The project is, per ETV Bharat, the first of its kind in India to wrap rip current detection, thermal imaging, weather monitoring and a public alert chain into a single deployable unit. Scientists say nearly 60 per cent of the AI software is already built, and a companion smartphone app called Safe Beach is expected within six months. The full details sit in the original Project Bharati partnership announcement.

What Project Bharati Actually Builds

Project Bharati is built around one piece of hardware: a solar-powered mobile safety unit that authorities can move from beach to beach as conditions change. Each unit costs around Rs 2 lakh and runs on its own onboard AI processor. The two-year project is being led by scientist Dr S.V.V. Arun Kumar of SAC-ISRO and Prof. C.V. Naidu, who heads the Department of Meteorology and Oceanography at Andhra University, with the same scope confirmed in the parallel two-year project write-up with the same team.

The hardware stack on each unit:

• AI-enabled 360-degree cameras tracking beach activity and wave patterns
• Thermal cameras for night-time and poor-visibility monitoring
• A weather station recording wind speed, wind direction, temperature, humidity and atmospheric pressure
• Warning sirens, flashing emergency lights and a public address system
• LED information boards displaying real-time safety messages
• A 4G/5G uplink to lifeguards, Coastal Security Police and control rooms
• Solar panels and a movable platform for relocation between beaches

Once a rip current is detected, the system auto-classifies sections of shoreline as safe, caution or danger zones. Alerts fan out simultaneously to visitors on the ground and to the agencies responsible for the response.

The Death Toll That Built the Case

Lifeguards under the Greater Visakhapatnam Municipal Corporation have rescued 199 people from drowning since December 2022, when the current 47-strong lifeguard force was set up, per the lifeguard rescue and drowning figures for Visakhapatnam.

The numbers paint a steady toll rather than a single crisis. Deaths peaked at 46 in 2022, with 35 in 2021, 33 in 2023, 27 in 2024, and 24 through 2025 to date, per Deccan Chronicle’s tally of city beach drownings. R.K. Beach alone has logged 42 drownings across the same five-year window. Bheemili recorded 15, Fishing Harbour 12, and Rushikonda and Yarada 10 each.

Two recent incidents underline how fast a beach day can turn. On October 5, 2025, 16 Russian tourists at Yarada Beach got caught in a rip current; lifeguards pulled all of them out, performed CPR on three, and saved two. One of the Russian tourists died. On September 21, 2024, five Italian tourists at the same beach were swept away and rescued.

By the numbers, Visakhapatnam’s beach safety baseline:

• 199 people rescued by lifeguards from 2023 through 2025 to date
• 46 drownings in 2022, the worst year on record in the recent tally
• 42 deaths at R.K. Beach across 2021 to 2025
• 165 drownings recorded across all city beaches from 2021 to 2025
• 47 lifeguards on duty across the coast under GVMC since December 2022

How the AI Spots a Rip Current

The detection layer pulls four streams at once. SAC’s AI software ingests satellite imagery, weather data, oceanographic information and live video feeds from the mobile units, then flags where and when a rip current is likely to form or has just appeared.

The scientists behind the project say rip currents along this coast can reach five metres per second and stretch 50 to 150 metres offshore, sometimes five to ten metres wide. They form most often in October, November, April and May, often where drainage channels and streams meet the sea, per the high-risk beach list and Safe Beach app design reported by ETV Bharat. SAC has been studying the phenomenon along the Visakhapatnam coast since 2017.

ISRO has named the beaches it considers highest risk. The pilot will track them first.

• Opposite Kali Mata Temple, Visakhapatnam
• Near Kursura Submarine Museum, Visakhapatnam
• Sai Priya Resorts area, Rushikonda, Visakhapatnam
• Yarada Beach sand dunes, Visakhapatnam
• Bheemunipatnam Beach river confluence, Visakhapatnam
• Suryalanka Beach, Bapatla district
• Maipadu Beach, Sri Potti Sriramulu Nellore district

Why This Is More Than a Local Pilot

It is a pilot, designed as a template, not a one-off installation. The same solar-powered mobile unit is meant to be drop-shipped to any Indian beach once the design holds up. The detection software sits on top of years of SAC field research along this coast. Project lead Dr S.V.V. Arun Kumar and his co-lead at Andhra University are building the system for replication, not for one beach.

Because these currents are often difficult for the public to identify, an early warning system can play a critical role in preventing accidents. This is a pilot project. Once successful, it is likely to be launched across the nation.

That quote is from Prof. C.V. Naidu, who heads the Department of Meteorology and Oceanography at Andhra University. Naidu’s framing matters because the same unit design is meant to travel to any Indian coastline, not just sit on Visakhapatnam sand. Naidu’s national-rollout prediction is the load-bearing line of the entire programme, and it is a forecast, not a fact.

SAC has been studying rip currents along the Visakhapatnam coast since 2017, and the partnership with Andhra University is how those years of research become a deployable safety tool. The same shoreline is being reshaped into a hyperscaler AI hub in parallel, with Alphabet’s $15 billion Google AI hub commitment placing the pilot in the middle of a much larger infrastructure build-out, as detailed in Visakhapatnam as a hyperscaler AI hub. If the unit survives salt, sun and salt-water spray, the same architecture can be replicated at any of India’s 7,500 km of coastline.

Where Project Bharati Could Stall

A 60 per cent software milestone, a sub-Rs 2 lakh unit price and a six-month app launch are real numbers, and they are also the project’s soft spots. The software has not been stress-tested through a full monsoon cycle. The unit’s solar-plus-AI power budget has to survive the Bay of Bengal’s humidity and salt spray. The 4G/5G alert chain only works where mobile coverage does, which is patchy on stretches of India’s coast. The software build, the hardware and the public-facing app are all still in development, per the project’s own progress report.

The other limit is institutional. The Coastal Security Police, GVMC, the Tourism Department and beach management authorities are the named beneficiaries, but no source has disclosed who runs the system after the pilot ends, who pays for fleet maintenance, or whether the Safe Beach app will require an active data connection at the moment a swimmer needs it most. Naidu’s national-rollout prediction stands as a forecast, and it depends on those answers landing first. The team has set a two-year clock; the test is whether the system clears it without losing support or coverage.

Frequently Asked Questions

What is Project Bharati?

Project Bharati is a two-year, Rs 22 lakh research programme run jointly by the Space Applications Centre of ISRO in Ahmedabad and Andhra University in Visakhapatnam. It is building an AI-enabled rip current detection and warning system for the city’s beaches, with a national rollout plan if the pilot succeeds.

How does the ISRO AI rip current detection system work?

The system pulls four data streams at once: satellite imagery, weather data, oceanographic inputs, and live video feeds from the mobile units on the beach. Its onboard AI then assigns each stretch of shoreline a risk level, from clear-water to red-zone, and triggers sirens, lights, public announcements, and 4G/5G alerts to lifeguards and the Coastal Security Police.

Which beaches in Andhra Pradesh will get the system first?

ISRO has flagged seven high-risk beaches across three districts for the pilot. The five in Visakhapatnam district are the Kali Mata Temple area, the Kursura Submarine Museum, Sai Priya Resorts at Rushikonda, the Yarada Beach sand dunes, and the Bheemunipatnam river-sea confluence. Suryalanka Beach in Bapatla and Maipadu Beach in Sri Potti Sriramulu Nellore round out the list.

When will the Safe Beach app launch?

Dr S.V.V. Arun Kumar of SAC-ISRO told ETV Bharat the app is expected within six months. It will let beachgoers scan a beach with their phone and see rip current zones highlighted in red, so they can move to a safer stretch before entering the water.

Is the AI rip current system already in use?

Not yet. Scientists involved in the project said nearly 60 per cent of the software is built. The mobile units, the Safe Beach app and the public alert chain are still in development or testing, and no source has confirmed a live deployment date on any of the named beaches.