Search and Rescue in Tunnels, an Integrated Approach Using Robotic Platforms

Abstract Underground infrastructure such as road and railway tunnels requires reliable communication, localisation, and situational awareness, particularly during search and rescue operations. This article summarises developments carried out within the A‑IQ Ready project in three closely related areas: underground wireless communication modeling, survivor detection, and robotic localisation. A tunnel-specific wireless simulation platform was developed and validated against measurements acquired at the underground facility Zentrum am Berg, enabling realistic prediction of quality-of-service metrics and supporting AI-based communication assessment. Measurement campaigns with a robotic platform were used to generate datasets and calibrate simulation models. For the detection of tunnel users or survivors, a multimodal thermal-LiDAR perception pipeline was implemented on the SPIDER unmanned ground vehicle (UGV). In parallel, LiDAR-based SLAM approaches were investigated for GNSS-denied tunnel environments, resulting in a centralised multi-robot mapping architecture based on LIO-SAM and map-merging concepts. Demonstrations at Zentrum am Berg confirmed the suitability of the developed technologies for search and rescue applications in complex underground infrastructure. They constitute key constituents of enhanced robotic assistance in underground emergency cases.