Periodic Reporting for period 1 - i-RASE (Intelligent Radiation Sensor Readout System)
Berichtszeitraum: 2024-03-01 bis 2025-02-28
The Intelligent Radiation Sensor System (i-RASE) project is a groundbreaking initiative dedicated to revolutionizing radiation detection and imaging technology. The primary goal of the project is to develop a scalable and sustainable sensor system-in-package (SIP) that can perform real-time (RT) radiation detection and imaging with unprecedented accuracy and speed. Current radiation sensor technologies are often hampered by slow data processing, low measurement accuracy, and high digital data output. By integrating advanced artificial neural network (ANN) technology, i-RASE aims to overcome these limitations and provide a versatile solution applicable across healthcare, environmental monitoring, space exploration, security, and defense.
The i-RASE project began with a kick-off meeting on 06 March 2024 at DTU Space. Top-level and work package (WP) requirements were presented. After six months, a technical and management meeting at POLIMI reviewed milestones and ensured alignment with objectives. Despite minor delays, the project is on track.
WP1: Developed and tested theoretical detector models and ANN for real-time event classification. Established a feedback loop between DTU Space and DTU Compute.
WP2: Defined requirements for the sensor signal converter and test ASIC. Prepared for system-in-package integration.
WP3: Compiled digital interface requirements and reviewed test setups. Integrated ANN with realistic inputs.
WP4: KROMEK proposed a budget change. Delivered the D-matrix system with pixel detectors. Advanced requirements for next-gen sensors.
WP5: Set up ANN test environments and conducted initial tests with real data. Validated ANN's ability to process detector signals.
Next Steps: Refining ANN architectures, initiating ASIC fabrication, finalizing the digital platform, integrating advanced sensors, and maintaining robust project management.
i-RASE aims to deliver a novel radiation sensor system using real-time AI for enhanced detection and imaging.
WP1: Developed and tested theoretical detector models and ANN for real-time event classification. Established a feedback loop between DTU Space and DTU Compute.
WP2: Defined requirements for the sensor signal converter and test ASIC. Prepared for system-in-package integration.
WP3: Compiled digital interface requirements and reviewed test setups. Integrated ANN with realistic inputs.
WP4: KROMEK proposed a budget change. Delivered the D-matrix system with pixel detectors. Advanced requirements for next-gen sensors.
WP5: Set up ANN test environments and conducted initial tests with real data. Validated ANN's ability to process detector signals.
Next Steps: Refining ANN architectures, initiating ASIC fabrication, finalizing the digital platform, integrating advanced sensors, and maintaining robust project management.
i-RASE aims to deliver a novel radiation sensor system using real-time AI for enhanced detection and imaging.
Real-Time ANN Integration
Embedded artificial neural networks for on-the-fly detector signal processing and event classification reduce data bottlenecks and surpass conventional offline analysis (cutting conventional event data size from ~50 KB/event to ~16 bytes/event for the 3D CZT DSD test detector). The system allows a direct connection to sensor signals for near-real-time, comprehensive signal processing, enabling the extraction of incident radiation information much faster and accurate.
Unified System-in-Package
Analog signal conversion, digital electronics, and AI are integrated into a single, compact, low-power system-in-package (SIP).
Physics-Informed Models
Synthetic data generation paired with real measurements accelerates robust ANN training and detector optimization. A potential IPR for a new ML method—combining physics-based parameters with real detector signals in an ANN-driven simulator—is under discussion.
Potential Impacts
Healthcare and Security: Faster, more accurate imaging and threat detection systems.
Environmental and Space: High-precision radiation monitoring with minimal hardware footprint, featuring intelligent on-board event processing software for background suppression and an improved signal-to-noise ratio based on radiation and interaction-type classification.
Key Needs for Further Uptake
Additional R&D and Demonstrations:
Explore cost-efficient detector materials to apply the i-RASE SIP approach for better detector performance and easier availability in cost-effective radiation detection and measurement applications.
Conduct ESRF test campaigns to confirm performance, reliability, and cost-effectiveness.
Access to Markets and Finance: Industry partnerships, private/public funding, and international collaborations.
Initial Proof of Concept Validated
Real-world 3D CZT DSD detector data confirm the feasibility of a simple trained ANN with enhanced speed and accuracy.
Prototype SIP
Early converter ASIC designs and digital platforms show promise for integrated, scalable radiation detection solutions.
Embedded artificial neural networks for on-the-fly detector signal processing and event classification reduce data bottlenecks and surpass conventional offline analysis (cutting conventional event data size from ~50 KB/event to ~16 bytes/event for the 3D CZT DSD test detector). The system allows a direct connection to sensor signals for near-real-time, comprehensive signal processing, enabling the extraction of incident radiation information much faster and accurate.
Unified System-in-Package
Analog signal conversion, digital electronics, and AI are integrated into a single, compact, low-power system-in-package (SIP).
Physics-Informed Models
Synthetic data generation paired with real measurements accelerates robust ANN training and detector optimization. A potential IPR for a new ML method—combining physics-based parameters with real detector signals in an ANN-driven simulator—is under discussion.
Potential Impacts
Healthcare and Security: Faster, more accurate imaging and threat detection systems.
Environmental and Space: High-precision radiation monitoring with minimal hardware footprint, featuring intelligent on-board event processing software for background suppression and an improved signal-to-noise ratio based on radiation and interaction-type classification.
Key Needs for Further Uptake
Additional R&D and Demonstrations:
Explore cost-efficient detector materials to apply the i-RASE SIP approach for better detector performance and easier availability in cost-effective radiation detection and measurement applications.
Conduct ESRF test campaigns to confirm performance, reliability, and cost-effectiveness.
Access to Markets and Finance: Industry partnerships, private/public funding, and international collaborations.
Initial Proof of Concept Validated
Real-world 3D CZT DSD detector data confirm the feasibility of a simple trained ANN with enhanced speed and accuracy.
Prototype SIP
Early converter ASIC designs and digital platforms show promise for integrated, scalable radiation detection solutions.