Progressive Field Deployment of Measurement-Constrained Robotic Systems for Radiological Investigation in Confined Nuclear Basins

Radiological characterization of legacy sludge basins is a critical issue for decommissioning, radiological safety assessment, and retrieval planning. Conventional inventories based on peripheral measurements and limited sampling are often insufficient when access is constrained and activity distributions are heterogeneous. This paper presents and validates an under-basin characterization methodology based on systematic ambient dose equivalent rate measurements acquired close to the source term under controlled geometry. The approach was implemented during two industrial campaigns on basins with increasing constraints (semi-accessible versus confined blind environment). A progressive deployment strategy was used to maintain metrological requirements, including controlled sensor-to-structure distance, stable detector orientation, and repeatable acquisition grids, while preserving operational robustness in contaminated and high-dose-rate conditions. Across both deployments, 5,168 (B1) and 2,909 (B2) measurement points were collected over five-day interventions, providing dense and spatially explicit radiological maps of accessible understructure areas. The campaigns demonstrate the feasibility of large-scale, geometry-constrained in situ data acquisition beneath concrete basins, with reduced operator exposure compared with direct human intervention. The results support improved radiological inventory refinement and provide a transferable methodological framework for characterization tasks in other constrained nuclear facilities.