Improving TLD personnel dosimetry of occupational workers

Thermoluminescence dosimeter (TLD) based personnel monitoring is widely employed for recording occupational radiation doses. TLDs used in routine monitoring are subjected to diverse field conditions, including processing and handling that result in variability in the glow curve (GC) shape. The distortions in the shape of GC may lead to variations in the estimated dose. The present work is focused on: (a) investigation and classification of anomalies in GC using machine learning algorithms, (b) development of an artificial intelligence (AI) based decision support system for screening GCs, and (c) providing a predictive maintenance system for the TLD badge reader. (author)