Intelligent Radiation Sensor Readout System

A single chip will process each particle of radiation as it arrives, turning a detector into an intelligent sensor that outputs data in real time. Today’s radiation detectors generate a flood of raw data that must be sent to a separate computer for analysis, which slows down imaging and consumes power. The i-RASE project embeds artificial neural networks—algorithms inspired by the brain’s wiring—directly into the sensor’s hardware. This allows the chip to classify and interpret each photon or particle on the fly, discarding noise and keeping only useful information. If the chip works as designed, it could shrink the size, cost, and energy use of radiation detectors while boosting their speed and accuracy. In medical imaging, that might mean lower radiation doses for patients and faster scans. In industrial inspection, it could enable real-time quality checks on production lines. Space instruments and environmental monitors would also benefit from a compact, low-power sensor that sends back only the data that matters. The work is primarily an engineering and fundamental science challenge—building a new class of sensor-system-on-a-chip that merges physics-based signal processing with neuromorphic computing. There is no immediate clinical or commercial product, but the underlying approach could reshape how any radiation-based system handles data.