← ResearchGUIDED / DEFENCE ROBOTICS

GUIDED: a clearer view for robot operators.

An Explosive Ordnance Disposal operator works at a distance from the scene. The robot's camera can show what is ahead, but understanding the surrounding space from a single video stream remains demanding.

GUIDED explores a different interface: a reconstructed 3D scene, useful cues from multiple sensors, and a view the operator can move through. Intermodalics develops localisation, integration and the VR operator interface within this Belgian defence research project.

Explore the project
Reconstructed outdoor scene with an EOD robot
01 / THE RESEARCH QUESTION

How do we turn sensor data into situational awareness?

More cameras do not automatically give an operator a better understanding of the scene. The information must be spatially consistent, relevant to the task and manageable during operation.

  • Understand the space around the robot. A 3D reconstruction can provide context beyond the viewpoint of a single camera.
  • Bring useful cues to the operator. AI-based analysis of multispectral images can highlight regions that deserve attention, including information outside the visible spectrum.
  • Keep the interface coherent. Scene geometry, live views and alerts need to work together without forcing the operator to continually switch between sensor feeds.

The research aims to improve situational awareness and reduce cognitive load while the operator remains in control.

02 / OUR CONTRIBUTION

From the robot's sensors to the operator's view.

Intermodalics connects the software components that acquire, localise and present information from the robot.

  1. Build the system architecture. We connect sensor acquisition and robot kinematics with mapping, reconstruction and analysis, then bridge that information to the visualisation interface.
  2. Localise across sensing modes. Our multi-modal SLAM work combines camera, inertial and other sensor information to locate the robot in challenging indoor and outdoor environments.
  3. Make the interface work as one system. We integrate the VR platform, live camera streams and 3D data, including operator interaction and alerts, and contribute to sensor-rig integration and demonstrator campaigns with DOVO.

Our partners bring complementary expertise: Trace works on real-time neural 3D reconstruction; the Royal Military Academy contributes end-user requirements, evaluation and multispectral salient-region detection.

Multi-modal SLAM map
Multi-modal SLAM map
VR operator interface with multiple camera streams
VR operator interface with multiple camera streams
03 / IN PRACTICE

Designed around the operator's task.

The setting. Two operational scenarios frame the work: a suspicious backpack in an indoor public space and a suspected explosive device hidden in an outdoor urban environment.

The research focus. A reconstructed scene supports a third-person or bird's-eye view. The operator can inspect spatial context while keeping access to live camera images and relevant sensor cues.

What we want to establish. Demonstrator work with DOVO is intended to test whether these capabilities support clearer, more intuitive operation. Feedback from the people controlling the robot guides the interface design.

Operator view of a reconstructed rubble scene
Operator view of a reconstructed rubble scene
04 / INTENDED OUTCOMES

A demonstrator that brings the information together.

The goal is an integrated demonstrator for a DOVO EOD robot, combining three capabilities.

  • Spatial context. Localisation and real-time reconstruction that help relate the robot to its surroundings.
  • Relevant visual cues. Multispectral analysis that draws attention to useful regions without requiring every sensor band to be watched continuously.
  • An integrated operator view. A VR interface that presents the scene, camera streams and alerts together.

The work builds on our robot localisation and perception engineering experience. Those same integration questions also arise in industrial robots operated at a distance.

VR-assisted robot operation in an outdoor demining scenario
VR-assisted robot operation in an outdoor demining scenario
05 / THE PARTNERSHIP

Robotics, reconstruction and defence research.

GUIDED stands for Goggles-based User Interface for Detection of Explosive Devices. Intermodalics works with Trace vzw and the Royal Military Academy, with input from DOVO end users.

GUIDED (DAP/23-08) is funded by the Belgian Defence-related Research Action (DEFRA), a programme of the Royal Higher Institute for Defence (RHID) and the Belgian Science Policy Office (BELSPO).

Read the public project description · Visit the Royal Military Academy project page

BELSPO
DEFRA
Belgian Defence
06 / YOUR NEXT STEP

Tell us about your defence or demining mission: inspecting suspicious objects, assessing explosive hazards, or keeping personnel at a safe distance. Let’s explore how we can support your teams in the field.

Let's discuss your defence or demining application.

Your Name

Please enter your email so we can get in touch.

Please enter your comments.