NAVIDUS

Autonomous Navigation of Robots in Darkness or Strongly Scattering Environments

• Reliable localization even in complete darkness or strongly scattering environments
• Autonomous navigation in GNSS-denied environments
• Integration of range-gated cameras and bio-inspired tactile sensors
• Multi-sensor fusion for terrestrial and underwater robots
• Intelligent environment perception and situation-aware trajectory planning

Reliable localization is a fundamental requirement for autonomous robots. However, conventional navigation systems can reach their limits when GNSS is unavailable and environmental conditions affect camera- or LiDAR-based perception. This applies particularly to operations in complete darkness, fog, heavy rain, snowfall, sandstorms or underwater environments with a high concentration of suspended particles.

The NAVIDUS research project addresses this challenge by developing a robust localization and navigation solution for autonomous robots. The core objective is to enable reliable localization even without GNSS and under difficult visibility conditions caused by darkness or strongly scattering light.

NAVIDUS therefore integrates two innovative sensor technologies: range-gated cameras and bio-inspired tactile sensors. Range-gated cameras combine short laser pulses with precisely synchronized camera exposure. This makes it possible to detect objects while significantly reducing the effects of scattered light. Tactile sensors inspired by animal whiskers provide additional information about physical contacts, surface structures and nearby obstacles.

Dedicated SLAM methods will be developed for both range-gated cameras and tactile sensors. These methods will be integrated into a comprehensive multi-sensor fusion solution combining inertial sensors, wheel odometry, conventional cameras, 3D LiDAR and, where available, GNSS RTK/PPP and echo-sounder measurements. The combination of complementary sensor technologies is intended to maintain precise and continuous localization even when individual sensors are unavailable or provide unreliable data.

In addition to localization, NAVIDUS will develop an intelligent environment perception system. The geometrical 3D map generated by SLAM will be enriched with semantic information. For terrestrial applications, the system will distinguish between traversable terrain, local obstacles, extensive hazardous terrain and acute danger zones such as steep slopes. For underwater applications, relevant objects include the seabed, pipelines, power and data cables as well as vessels showing potentially suspicious behavior.

Based on this intelligent environment model, autonomous robots will be able to select suitable routes and maneuvers according to the current situation. Trajectory planning will take into account risks such as collisions, sinking into soft ground, becoming trapped in dense vegetation or falling from steep terrain. Robust and situation-adaptive control methods will further increase the reliability of the autonomous system.

ANavS leads the NAVIDUS project and is responsible for precise localization, multi-sensor fusion and intelligent environment perception. Schmalkalden University of Applied Sciences contributes its expertise in route and trajectory planning, decision-making, tactile sensing and the development and navigation of terrestrial and underwater robotic platforms.

The developed technologies will be validated in realistic terrestrial and underwater test scenarios. Potential future applications include autonomous exploration robots, monitoring of hazardous areas, protection of critical underwater infrastructure, agricultural robotics, autonomous driving and space exploration.