Nantista project is composed of five Work-Packages (WPs). The timeline of WPs is presented in figure below.
Each Work-Package (WP) specifically addresses project problematics and is managed by a member of the consortium.
Workpackage is subdivided in tasks and each task is managed by a member of the project.
WP0 Project management
This WP deals with project management, project coordination and communication. The project
coordinator is CEA LIST. It comprises only one task for a total duration of 42 months.
- T0.1 Project coordination
WP1 Radionuclide database and experiments
Radionuclides are characterized by a specic decay-scheme depending on several features: ionizing emissions, energy spectrum (discrete or continuous), half-life, etc. Therefore, the learning procedure required for ANN processing has to be implemented using specic database for each radionuclide iden-tied as a potential radiological threats. The main purpose of this WP is to constitute a library of raw data made of digitized pulses obtained with a nuclear instrumentation using radioactive sources prepared at LNHB. As some specic radionuclide cannot be handled in safe condition, Monte Carlo simulations knowing the detector response and decay-schemes available in the literature can provide synthetic signals. This WP will be managed by LNHB. It has a long-term experience in radiochemistry and source preparation for radioactivity measurements. It has the authorization to handle all radionuclides used for instance in nuclear industry or for medical applications. The laboratory is also experienced in modeling nuclear detectors for Monte Carlo simulations and in the development of specialized modules for radioactivity metrology. Both LCAE and SAPHYMO will be involved in that WP by providing detectors and electronic chain characteristics and realistic test-case respectively. This WP comprises three tasks and is managed by CEA.
- T1.1 test-cases and mixtures of radionuclide denition
- T1.2 experimental data acquisition with various radionuclide
- T1.3 database constitution from various simulations
WP2 Signal processing and ANN architecture design
This WP consists in dening best topologies for ANN for neutron-gamma discrimination and radionuclide identication. Preliminary results on that topic have lead to a patent that is currently in progress. This WP relies on radionuclide signal acquisition database and simulation. Thanks to previous work done by CEA, a pulse acquisition system was designed which it makes possible to eciently acquire individual pulses. The system is patented [Comment faire lien REF 34 ?] and is a prerequisite to our proposal. This database will be used as learning model for neural network architecture. This WP will benet from the experience of consortium members in terms of signal processing, neutron-gamma discrimination methods and radionuclide identication. As described above in the Introduction, preliminary results were obtained by CEA in advance. This WP comprises three tasks.
- T2.1 ANN theory and design for neutron-gamma discrimination
- T2.2 ANN theory and design for radionuclide identication
- T2.3 signal preprocessing for ANN
WP3 Prototyping and pre-industrialization
Once the theorical approach is investigated in the previous WP, this one aims at developing a complete system that is able to embed our proposed approach. The complete system will comprise a mezzanine acquisition board associated with an on-the-shelf digital signal processing board. The acquisition board will have to pre-process the signal as dened in taks 1.3, and extract pulses informations. This WP will be managed by CEA. The deliverable of the WP is a complete prototype that is able to detect threats dened in test cases. WP3 comprises five tasks and is managed by CEA.
- T3.1 acquisition board design
- T3.2 signal pre-processing rmware design
- T3.3 ANN implementation
- T3.4 real-condition test scenario
- T3.5 pre-industrialization study
WP4 Results dissemination
This WP focuses on project development valorisation and dissemination of results. While traditional dissemination of results of the project is ensured by patenting and publishing in journal and conferences and undergraduate students courses, a specic task consists in the creation of an on-line database to incite signal processing community to challenge their solution on our dataset to solve the problem. Vulgarization is ensured by specically selected publications thanks to CEA communication department. It comprises three tasks.
- T4.1 patenting and licensing
- T4.2 publication and courses
- T4.3 on-line database publication