The SilentBorder project has advanced the development of cosmic ray tomography as a powerful complementary tool to X-ray scanning for detecting hazardous and illegal goods in trucks and sea containers.
A major technical achievement was the design and validation of the hodoscope module, the core detector element comprising muon-sensitive scintillation fibers. A prototype was built and tested successfully, proving the feasibility of the production technology.
Alongside this, the Data Acquisition System (DAQ) was developed and tested. This is especially complex due to the large number of fibers in each hodoscope.Future R&D for such large-channel systems was mapped out.
DAQ development was accompanied by integration of a special nuclear materials detection solution - the SNIPER - into the cosmic-ray tomography technology.
To guide system optimization, a comprehensive Monte Carlo modelling framework was established to simulate muon transport and detection in a realistic muon tomography system. A detailed computer model was developed to describe the muon tomography system and different cargoes, evaluate the performance of the cosmic-ray tomograph and suggest ways to optimize detectors and analysis methods. Many scenarios have been simulated, the data and images stored for future use as templates for material identification. A new code TomOpt was developed, based on simplified simulations and a novel programming technique (differentiable programming), to optimize muon tomography system designs. An assessment of the parameters of the designed muon tomography system using simulated data revealed a few improvements that could enhance its performance and lower costs. These improvements may require further development of detectors and analysis methods, in particular, using machine learning techniques such as neural networks.
A core part of the SilentBorder project was the development and optimization of the tomography system as a whole. A physical tomography system was designed, incorporating algorithms that facilitate tomographic measurements. Key achievements include the development and initial validation of tomographic reconstruction methods using machine learning, and classification methods that have demonstrated over 90% accuracy with simulated data for simple shape detection and material classification. Technical design documentation and drawings, as well as a report on the system's relocatability and flexibility were produced. Software for detector tracking, alignment, and data analysis was also developed.
The SilentBorder system was tested and demonstrated to the public on three separate occasions.
The 1st demonstration focused on providing the basic functionality of the system. Small scale prototype successfully detected muons and produced images. The early version also integrated AI algorithms to help interpret the data, demonstrating how artificial intelligence can enhance imaging even in the project's initial stages.
In the second demonstration, the SilentBorder technology was shown to be fully compatible with the SNIPER special nuclear material detection system developed by CAEN. This highlighted the system's flexibility and its potential to work alongside existing nuclear security tools.
The 3rd and final demonstration took place in a realistic outdoor environment. A fullscale prototype was tested together with X-ray scanning technology. By innovative AI methods, the system successfully identified materials missed by X-rays alone in under 50 minutes.
Together, these results demonstrate SilentBorder’s strong potential to provide a safe, effective, and non-invasive inspection method for border security. By combining cosmic ray tomography with advanced AI-driven analysis, the project has opened the way towards future systems that can reliably detect concealed materials and support customs and law enforcement agencies in protecting society. In addition to advancements in identifying low atomic mass materials, the integration of the patented SNIPER system means that SilentBorder will also push boundaries in the detection and identification of special nuclear materials.
