OW
Menu
02 / TECHNOLOGY WITH A PURPOSE

Fire. Safety.
Rescue.

More awareness.
Better decisions. A coordinated response.

SCROLL TO EXPLORE
Illustrative industrial-safety drone concept
AN INDUSTRIAL SAFETY CONCEPT

See the heat. Before the hazard.

Follow an illustrative drone patrol from launch to thermal observation, a closer look by a ground robot and a shared site plan.

Illustrative industrial-safety concept. Proposed capabilities are untested; this is not a deployed system.

Original film · Case study
Read the story

Follow an illustrative drone patrol from launch to thermal observation, a closer look by a ground robot and a shared site plan.

  1. Launch on a schedule or respond to an alarm.
  2. Observe a possible hot spot from the air.
  3. Assign a ground robot to take a closer look.
  4. Bring observations into one record and coordinate a site plan.
FROM EARLY AWARENESS TO ASSISTANCE

Built around
what matters.

Fire

Exploring aerial observation and early-warning applications to help teams understand a developing incident.

Safety

Research into navigation, obstacle avoidance and decisions in places that are difficult to access.

Rescue

Cooperating drones that map an unfamiliar environment, share tasks and support the delivery of aid.

01 / CITY, LONDON

Go inside
the lab.

Real drones. Robotics. The space where ideas become experiments.

Step into the robotics and AI laboratories at City St George’s, University of London, in our short film.

01:09 / LAB FILM
Inside the City robotics and AI labsOpen video ↗ (new tab)
OW ROQUAI / ENERGY-SITE CONCEPT

One watch.
Air, ground, sound.

An energy-site concept bringing aerial observation, ground robotics and acoustic sensing into one proposed approach.

3:31 · Concept film with stock footage and simulation. The response sequence at 2:01–3:12 is simulated; the film does not demonstrate validated OW capabilities.

Still from the OW ROQUAI energy-site concept film
Open video : One watch: air, ground, sound (new tab)
02 / SAPIENCE · NATO SCIENCE FOR PEACE AND SECURITYCOMPETITION WINNER / 2024

First place.
Real drones.
Shared intelligence.

At City’s 2024 search-and-rescue competition, the university team’s two drones turned research into a winning real-world demonstration.

Our team’s research experience includes SAPIENCE: autonomous navigation and cooperation in a simulated disaster environment, without GPS.

Research quadrotor with labelled companion processor, RGB-D camera and LiDAR sensor
Research platform · Published experiment
Figure 3 · Paper & full attribution ↗ · CC BY 4.0
1stCity’s team
London event · 2024
€1.2mNATO SPS funding
Whole SAPIENCE programme
4University teams
One collaborative programme
THE WINNING RESEARCH / JOURNAL OF FIELD ROBOTICS · 2026

Deep Reinforcement Learning Based Autonomous Decision-Making for Cooperative Uncrewed Aerial Vehicles: A Search and Rescue Real World Application

Two physical drones combine deep reinforcement learning, shared task allocation and LiDAR-based navigation to map an indoor disaster scenario and coordinate delivery tasks without satellite positioning. The study documents City’s first-place result at the 2024 NATO SAPIENCE competition.

Read the open-access paper

The funding supported the four-university programme. The competition result belongs to City’s research team and is separate from OW product validation.

03 / EXPLORE THE TECHNOLOGY

Sense navigate work together

Scroll through the methods behind cooperative search and rescue. Two drones, one shared mission.

SCROLL TO EXPLORE
Published research photograph of the indoor drone test arena
THE METHODS IN MOTION

Sense navigate work together

Follow two research drones as they sense an indoor environment, navigate around obstacles and work towards a shared mission.

Simplified illustration of published SAPIENCE methods, not a recorded flight or validation of OW products.

Original explorer · Case study
Read the story

Follow two research drones as they sense an indoor environment, navigate around obstacles and work towards a shared mission.

  1. Sense: LiDAR, depth and inertial measurements build a picture of the surroundings.
  2. Navigate: learned guidance helps find a route through obstacles.
  3. Work together: task allocation connects mapping, search and aid delivery.
  4. Explore the physical arena and published research behind the illustration.

Published method and figure attribution ↗ · CC BY 4.0

04 / OW HARDWARE & DEVELOPMENT

From the inside.
Built to explore.

Our photographed prototype, the electronics inside it and the digital assembly. A closer look at the technology we are developing.

The original OW R1 quadrotor prototype with exposed electronics, four propellers and landing legs
01 / THE ORIGINAL BUILDR-300 Faraday Cage lineageExplore the prototype & concept ↗

Original hardware photography. The Faraday Cage concept and digital assembly remain in development; published research results are not product specifications.

SEE THE METHODS IN MOTION

A path through
the challenge.

Watch a UAV trajectory develop through a dense obstacle field in the S³F navigation simulation.

00:29 / RESEARCH SIMULATIONExplore the navigation project ↗
UAV obstacle navigation — S³F simulationOpen video ↗ (new tab)
LABS & ACADEMIC CONNECTIONS

Research with
reach.

Laboratory work and academic connections underpin a wider programme of robotics, control and autonomous-systems research.

LONDON / CITY ST GEORGE’S

From simulation
to physical systems.

Our team’s work at City connects robotics, autonomous flight and laboratory testing. OW has supplied a drone to the university; the laboratories are university-owned, with project access arranged.

Explore the robotics facilities ↗
MIT / RESEARCH AFFILIATION

A wider academic
connection.

Our team includes an MIT research affiliate, with research connections to the Mechatronics Research Laboratory. An academic reference from an MIT professor is available on request.

Request an academic reference ↗
THE WIDER BODY OF WORK

Go deeper.

Planetary robotics. Explainable AI. Robust control. Explore the papers behind the experience.

01IEEE International Conference on Robotics and Automation · 2021 / Conference paperAutonomous Cooperative Visual Navigation for Planetary Exploration Robots

Robots share visual observations to improve localisation where GPS is unavailable. The approach was evaluated using a planetary analogue dataset within the PRO-ACT programme.

Read the paper ↗
02Acta Astronautica · 2022 / Journal paperRobust deep learning LiDAR-based pose estimation for autonomous space landers

Deep learning and 3D LiDAR estimate position and orientation during landing. The study combines simulated planetary terrain with physical aerial-robot experiments.

Read the paper ↗
03ACM Computing Surveys · 2024 / Journal paperExplainability in Deep Reinforcement Learning: A Review into Current Methods and Applications

A review of methods that make learned decisions easier to understand, covering applications in navigation, control and robotics.

Read the paper ↗
04International Journal of Robust and Nonlinear Control · 2026 / Journal paperValidation of Neural Network Controllers for Uncertain Systems Through Keep-Close Approach: Robustness Analysis and Safety Verification

Research into verifying neural-network controllers against uncertainty, developed through the ESA-supported AITIVE-GNC programme.

Read the paper ↗
05International Symposium on Artificial Intelligence, Robotics and Automation in Space · 2020 / Conference paperMulti-Robot Cooperation for Lunar Base Assembly and Construction

The PRO-ACT programme brings different robots together for lunar assembly and construction research, with shared software, hardware and experimental test environments.

Read the paper ↗
06arXiv · City–MIT research · 2026 / PreprintA Fast and High-Accuracy Finite-Time Sliding-Mode Algorithm for Computing Local Stackelberg Equilibria in Nonlinear Bilevel Games

A method for solving nonlinear leader–follower optimisation problems, studied through mathematical analysis and numerical experiments. This work is available as a preprint.

Read the preprint ↗

Earlier City projects include Dstl-funded research into explainable AI and adversarial learning, ESA guidance and navigation work, and PRO-ACT planetary robotics.

CURRENT WORK / MULTI-ROBOT SIMULATION

A simulation of robots assigning and reassigning tasks as conditions change, accounting for deadlines, battery levels and task grouping. Physical experimental validation remains a next step.

Watch the task-allocation demonstration ↗
Planning your fire, safety & rescue project

Discuss the application, research stage and operating environment with us. OW drones remain in development; fire and rescue use requires project-specific testing. Academic experiments and simulations are not validated OW product specifications.

BUILT TOGETHER

Give your next idea lift.

Let’s build what comes next.

Start a conversation