Design, control and implementation of AI-driven autonomous snake robot for archeological site exploration

Abstract

Abstract For decades, archeological site exploration has presented major challenges due to narrow spaces, unstable structures, and the high sensitivity of historical artifacts. Recent advancements in robotic technologies have introduced highly capable and reliable alternatives for safe and efficient exploration in environments that are otherwise inaccessible or hazardous to humans. This paper presents mathematical modeling, design, control and implementation of a bio-inspired autonomous snake robot tailored for the exploration of confined environments, with a particular focus on archeological sites. Our primary contribution is a novel system integration approach designed specifically for highly confined and fragile heritage environments, successfully combining a modular alternating-joint architecture with a ROS-based control framework and an AI-driven perception system focused strictly on YOLOv11-based artifact and obstacle identification. A physical network model was developed using MATLAB Simscape to simulate the system dynamic behavior under varying payloads. The control architecture was implemented using the robot operating system (ROS), supporting real-time execution of multiple locomotion strategies through coordinated servo actuation. The system is validated through simulation and physical experiments, demonstrating stable undulatory locomotion, accurate mapping in confined environments, and reliable artifact detection under variable lighting and occlusion. The results indicate that snake-inspired robotic platforms provide a viable and scalable solution for autonomous exploration and documentation in sensitive heritage sites.