REISAR Underground Dataset with Ground Truth – Test of July & September 2025

Project Context: REISAR & France 2030 This work is part of the REISAR project (Advanced Robotic System for Sewer Network Inspection and Water Preservation), funded under the ANR-23-DMRO-0014 reference and supported by the France 2030 investment plan. The REISAR initiative addresses major challenges in clean water access and sustainable resource management, in alignment with the United Nations' Sustainable Development Goals for 2030. Sewer networks, essential for wastewater transport, are often aging, difficult to access, and hazardous for maintenance personnel. Starting in 2026, French regulations will require precise mapping (accuracy < 40 cm) of underground structures in sensitive areas, expanding to all urban zones by 2030. This project contributes directly to meeting these standards by validating high-precision localization technologies. REISAR proposes an innovative robotic solution: a ground-based intelligent robot capable of navigating and inspecting sewer tunnels with minimal visual features. The project explores: Robust localization in low-texture environments Semantic data transmission from robots User-friendly interfaces for non-experts Training and operational safety improvements The consortium(Pilgrim Technology , TRAAK , Conscience Robotics), ensuring a multidisciplinary approach. The “Made in France” nature of the solution highlights national technological excellence and its adaptability to other demanding environments. REISAR Underground Dataset with Ground Truth – July & September 2025 The REISAR Underground Dataset – July & September 2025 was collected as part of the REISAR project under the Défi Transfert Robotique initiative, a collaboration between CESI LINEACT, Pilgrim Technology, and other industrial and academic partners. This dataset supports research in localization, sensor fusion, and autonomous navigation in underground-like environments. The July 2025 recordings focus on validating multi-sensor integration and system synchronization in controlled indoor and outdoor settings. These tests served as a preparatory benchmark to ensure reliable data acquisition from RGB-D cameras, LiDAR, IMU, and ground truth systems such as the Trimble SX12 and TRAAK Wheere. The newly added September 2025 ROS bags introduce more advanced and realistic scenarios, recorded inside underground pipes made of different materials, including PRV (Polyester Renforcé de Fibres de Verre) and concrete. These tests simulate operational conditions such as confined spaces, curved tunnels, and material-specific challenges, offering a richer benchmark for evaluating navigation robustness, sensor fusion algorithms, and terrain adaptation strategies.