RadScopeX

Radiation Monitoring System
(AI-Based 3D Radiation Mapping System)

Brochure

International PCT Patent Application Filed 2 Domestic Core Patents Filed

Size	N/A (Embedded Software)
Weight	N/A	HS Code	8523.49 / 9030.10
Key Specifications	○ Core Algorithm: Physics-Informed Neural Network (PINN) + Radiation Transport Physics ○ Input Data: Sparse Gamma Ray Count Rate + 3D LiDAR / SLAM Spatial Coordinates ○ Output Data: 3D Radiation Voxel Distribution Map & Uncertainty Quantification (UQ, 95% CI) ○ Processing Speed: < 10 minutes post-survey completion ○ Hardware Compatibility: NVIDIA Jetson Orin / AGX Edge Modules, ARM/x86 Linux OS
Application Areas	○ Nuclear Power Plant Decommissioning & Facility Characterization ○ Nuclear & Radiological Emergency Response (CBRN Defense) ○ Facility Environmental Safety & Radioactive Waste Management

○ RadScopeX is a breakthrough onboard physics-informed AI edge engine designed to reconstruct high-precision 3D radioactive contamination maps in real-time using sparse data collected by autonomous vehicles such as drones and robots. - Zero-Field-Training via PINN: By directly embedding radiation transport physics equations into neural networks, it enables instant high-precision 3D mapping without any prior on-site training data. - Real-Time Onboard Edge Processing: Optimized for low-power edge hardware (e.g., NVIDIA Jetson), delivering complete 3D spatial mapping within 10 minutes post-survey. - Regulatory-Grade Uncertainty Quantification (UQ): Provides 95% confidence bounds essential for nuclear regulatory validation and worker exposure optimization. - Seamless Unmanned Integration: Effortlessly interfaces with 3D LiDAR, SLAM positional data, and radiation sensors on autonomous platforms.