A fast and reliable algorithm for the detection of low-radioactivity moving sources based upon an adaptive filter and a MACD estimator
Résumé
This paper introduces a novel method for detecting low-intensity changes in radiological activity regimes, which has been under development at CEA List since 2018. The method operates under a signal-to-noise ratio of approximately four in both numerical and experimental studies, with optimization and principle validation detailed herein. The procedure, which forgoes the use of radioisotopic spectral signatures in a controlled environment to meet the constraint of a reliable response time of approximately one second, primarily utilizes a real-time variation indicator. This indicator is analogous to the convergence-divergence of moving averages commonly employed in financial analysis. Furthermore, the article presents evidence from numerical simulations and experiments that this indicator yields figures of merit that are competitive with the state of the art when evaluated on a counting signal pre-filtered by an adaptive filter. This approach proves advantageous for two primary reasons: (i) it is more readily parametrizable, and (ii) its resilience against deviations between experimental nuclear counts and ideally modeled events drawn from Poisson distribution statistics is significantly enhanced.
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|---|
| Licence |