Adaptive Entropy-Driven Sensor Selection in a Camera-LiDAR Particle Filter for Single-Vessel Tracking
arXiv:2603.08457v2 Announce Type: replace-cross Abstract: Robust single-vessel tracking from fixed coastal platforms is hindered by modality-specific degradations: cameras suffer from illumination and visual clutter, while LiDAR performance drops with range and intermittent returns. We present a particle-filter tracker that supports sequential measurement-level camera-LiDAR fusion and an information-gain (entropy-reduction) adaptive sensing policy that selects the most informative sensing modality at each fusion time bin. The approach is validated in a real maritime deployment at the Cyprus Ma
The increasing sophistication of multi-modal AI systems and the urgent need for robust surveillance, particularly in maritime security, drive the development of adaptive sensor fusion techniques.
This development allows for more reliable and resilient autonomous monitoring and tracking systems in challenging environments, significantly improving situational awareness and operational effectiveness for defence and security applications.
The ability to dynamically select the most informative sensing modality in real-time enhances the robustness of tracking systems against environmental degradations, moving towards more autonomous and less human-dependent surveillance.
- · Defence contractors
- · Maritime surveillance operators
- · AI/robotics companies
- · Coastal security forces
- · Legacy un-fused sensor systems
- · Manual surveillance methods
Improved accuracy and reliability of autonomous vessel tracking in complex coastal environments.
Reduced operational costs and increased efficiency for maritime security, enabling quicker threat detection and response.
Potential for broader implementation of adaptive multi-modal sensing in other critical infrastructure monitoring and autonomous navigation systems.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG