Skip to content

Before humans enterdangerous waters,intelligence goes first.

01The Problem

When cities flood,responders enterthe unknown.

Floodwater can hide obstacles, changing currents and isolated people, while emergency teams often operate with incomplete information.

  • 01Flood environmentUnlit. Unmapped.
  • 02Search radiusArea expands with time.
  • 03Thermal returnsPossible persons.
  • 04Obstacle fieldSubmerged hazards.
  • 05CommunicationRelay to response teams.
  • 01Search

    Search is difficult

    Floodwater creates large, obstructed and constantly changing search areas.

  • 02Exposure

    Responders are exposed

    Teams may enter uncertain water before knowing what lies ahead.

  • 03Fragmentation

    Existing tools are fragmented

    Different tools can provide mobility, aerial visibility or sensing, but rescue operations often require those capabilities to work together.

Before rescuers enter the unknown,ARTICUNO R goes first.

ARTICUNO R three-quarter view showing the sensor turret, hull and rear payload bay
02The Machine

IntroducingARTICUNO R

A compact autonomous surface robot engineered for disaster-response operations.

Development platform
Reel — continued00:03 — 00:10
ARTICUNO R three-quarter view showing the sensor turret, hull and rear payload bay
Technical drawings

Insidethe machine.

Engineering visualisations of the development platform — internal arrangement, assembly order and the industrial-design study behind the hull.

03Mission Flow

One platform.One continuousrescue loop.

A single autonomy core carries the vehicle through every stage of a response, from launch to recovery.

Entry point
01Stage

Deploy

Place the unmanned platform into the response area.

Two-person launch from bank, boat or bridge. No dedicated ramp required.

Coverage path
02Stage

Navigate

Travel through the assigned search area using onboard navigation.

Route generated from the assigned survey boundary and held under GNSS + IMU.

Continuous sweep
03Stage

Search

Continuously scan the surrounding environment.

Sweep pattern maintained while ranging and imaging run in parallel.

Multi-modal correlation
04Stage

Detect

Use multiple sensing modes to identify possible people and hazards.

RGB and thermal returns are correlated before anything is surfaced as a candidate.

Operator verifies
05Stage

Verify

Transmit location and visual context to response teams.

A human decides. The platform supplies position, imagery and confidence.

Payload on authority
06Stage

Assist

Deliver mission-configurable emergency assistance.

Payload bay approaches the verified position and releases on operator authority.

Recovery reserved
07Stage

Return

Return to the defined point or continue the assigned mission.

Recovery path is held in reserve throughout the mission, not computed at the end.

04Perception

The water doesn'tlook the sameto every sensor.

So we don't relyon just one.

ARTICUNO R is being engineered around complementary sensing modes, so a failure or blind spot in one channel does not blind the platform.

Sensor renderings on this page are illustrative representations of intended perception behaviour. They are not captures from a validated production system.

05Autonomy

Autonomy is morethan followingwaypoints.

ARTICUNO R is being engineered around mission-level autonomy designed for unpredictable environments.

06Failure-Aware Intelligence

Real environmentsdon't followthe plan.

The system must adapt.

Degradation is treated as an expected mission condition. Each fault class is designed to have a defined detection path and a defined response, rather than a single catch-all failure mode.

  1. Detect
  2. Assess
  3. Adapt
  4. Continue / Return
  • F-01
    Nominal

    GNSS degraded

    Positioning confidence drops near structures and dense cover.

    DetectAssessAdaptContinue / Return

    Designed to fall back to inertial dead-reckoning and tighten the survey pattern until the fix recovers.

  • F-02
    Nominal

    4G unavailable

    Cellular coverage is often the first casualty of a disaster.

    DetectAssessAdaptContinue / Return

    Architecture supports an RF fallback link, with the mission continuing under last-authorised objectives.

  • F-03
    Nominal

    Sensor obstructed

    Spray, debris and rain can occlude an imaging channel.

    DetectAssessAdaptContinue / Return

    Intended to down-weight the affected channel and continue on the remaining sensing modes.

  • F-04
    Nominal

    Low battery

    Endurance margin is consumed faster in current and chop.

    DetectAssessAdaptContinue / Return

    Return energy is reserved throughout the mission, so recovery is not a late calculation.

  • F-05
    Nominal

    Propulsion degraded

    Debris fouling and thruster loss change vehicle handling.

    DetectAssessAdaptContinue / Return

    Designed to re-trim available thrust, reduce the objective and hold a recoverable heading.

  • F-06
    Nominal

    AI uncertainty

    A detection may be ambiguous or contradicted between sensors.

    DetectAssessAdaptContinue / Return

    Low-confidence returns are escalated to the operator rather than acted on autonomously.

The behaviours described are system design goals for the platform architecture, not validated field performance.

07System Architecture

Intelligenceat the edge.

Perception and mission decisions are designed to run onboard. The link to the command centre carries intent and evidence — not the vehicle's ability to keep itself safe.

Key hardware & systems

  • Sensors

    RGB camera, thermal camera, 360° LiDAR, ranging sensors, GNSS, IMU.

  • Compute

    NVIDIA Jetson Orin / NX class edge computing for onboard inference.

  • Autopilot

    Pixhawk-class autopilot for navigation and vehicle control.

  • Communication

    4G / 5G modem with RF radio for long-range telemetry and fallback.

  • Power system

    Smart battery with BMS and protected power distribution.

  • Propulsion

    Dual water-jet thrusters for high-efficiency, precise control.

  • Safety systems

    Independent safety MCU, watchdog, leak detection and emergency stop.

08Safety Architecture

Autonomymust failsafely.

Safety is designed as an independent, layered system rather than a single software decision.

  • 01

    Mission Intelligence

    Objective-level reasoning. Decides what the vehicle should attempt, and when to escalate to an operator.

    • Objective state
    • Confidence handling
    • Operator escalation
  • 02

    Vehicle Control

    Motion-level control. Converts mission intent into heading, thrust and station keeping.

    • Autopilot
    • Attitude control
    • Envelope limits
  • 03

    Independent Safety Controller

    A separate microcontroller outside the AI stack, designed to hold authority over the propulsion power path.

    • Watchdog
    • Link-loss handling
    • Independent cut authority
    Outside the AI stack
  • 04

    Physical Safety

    Hardware that does not depend on software being correct.

    • Emergency stop
    • Main contactor
    • Circuit protection

No single AI computershould have sole authorityover vehicle safety.

09Capabilities

Built aroundthe mission.

Each subsystem is scoped to a specific job in the rescue loop, and is being developed to work with the others rather than in isolation.

Holds a line through open and obstructed water on its own.

  • 01

    Autonomous Navigation

    GNSS + IMU driven navigation architecture.

    Design intent

    Designed to hold an assigned survey pattern and recover its route without continuous operator control.

  • 04

    Edge Intelligence

    Onboard mission processing with ROS2-based architecture.

    Design intent

    Perception and mission logic run on the vehicle, so autonomy survives a degraded link.

22kgWeight
25km/hMax speed
5kmOperating range
6hrsBattery life

Design targets for the development platform. Figures are engineering intent and have not yet been validated in the field.

Modular payload

Plug.Lock.Mission ready.

The rear bay is being designed so mission payloads can be exchanged without redesigning the vehicle around them.

  • 01Rescue ring deployerFlotation delivered to a verified position.
  • 02Medical kit moduleSealed supplies for an isolated casualty.
  • 03Public addressVoice contact before physical contact.
  • 04High-intensity lightScene illumination for night operations.
  • 05Water sample collectionContamination sampling in hazardous water.
  • 06Sonar / multi-beamSub-surface survey and obstruction mapping.
10Mission Environments

Where conditionsare dangerous,autonomy can go first.

The platform is being developed for flood and disaster response first. The same autonomy core is intended to extend to adjacent water environments where sending a crew is slow, costly or unsafe.

Drag to explore

01Primary

Flood Response

Urban and rural flooding where the search area is large, obstructed and changing hour to hour.

  • Primary mission
  • Urban water
  • Search + assist
01 / 05
02

Dam & Reservoir Response

Controlled water bodies where release events and steep banks make crewed approach hazardous.

  • Incident response
  • Restricted access
02 / 05
03

Industrial Water Emergencies

Process water, settling ponds and contained sites where exposure risk is chemical as well as physical.

  • Hazard survey
  • Contained water
03 / 05
04

Port & Harbour Operations

Congested working water where routine survey competes with vessel movements.

  • Survey
  • Congested water
04 / 05
05

Maritime Monitoring

Persistent nearshore observation tasks that are repetitive, long and poorly suited to a crew.

  • Persistent tasking
  • Nearshore
05 / 05
11Engineering Philosophy

Reliabilitybeforecomplexity.

ARTICUNO R is being developed through a staged test–validate–evolve approach. Each layer has to hold on real water before the next one is added on top of it.

  1. 01Phase

    Mobility

    Prove the platform on water first — hull behaviour, propulsion, power, thermal load and manual control.

    • Hull & propulsion
    • Power and endurance
    • Operator control
  2. 02Phase

    Autonomy

    Add navigation and mission execution once the vehicle itself is predictable and recoverable.

    • GNSS + IMU navigation
    • Route holding
    • Failure handling
  3. 03Phase

    Rescue Intelligence

    Layer perception, detection and assistance behaviour onto a platform that is already dependable.

    • Multi-modal perception
    • Detection & verification
    • Payload assistance
Working rules
  • 01

    Test on water, not only in simulation.

  • 02

    A degraded mission is better than a lost vehicle.

  • 03

    Safety authority stays outside the AI stack.

  • 04

    Operators verify. The platform informs.

12Platform Vision

One autonomy core.Multiple missions.

The navigation, perception and safety architecture being built for rescue is intended to carry across to other water environments. Configurations beyond ARTICUNO R are directions of development, not available products.

  • ARTICUNO R — Rescue configuration
    RescueIn development

    ARTICUNO R

    The development platform. Flood and disaster-response missions, built around search, verification and assistance.

  • ARTICUNO Industrial — Inspection configuration
    Future platform
    InspectionFuture platform

    ARTICUNO Industrial

    A direction of development for contained industrial water — survey and inspection tasking on the same autonomy core.

  • ARTICUNO Maritime — Monitoring configuration
    Future platform
    MonitoringFuture platform

    ARTICUNO Maritime

    A direction of development for persistent nearshore monitoring, extending endurance and tasking rather than payload.

ARTICUNO R travelling across open water at dusk
13Contact

When everysecond counts,go unmanned first.

Autonomous.Intelligent.Life saving.

15Our Team

The peoplebuilding it.

A small team building ARTICUNO R end to end — hull, electronics and the autonomy stack.

  • Somesh Choudhary, Founder at ARTICUNO R

    Somesh Choudhary

    Founder
  • Harsh Kamble, Tech Lead at ARTICUNO R

    Harsh Kamble

    Tech Lead