FPGA-based singles and coincidences processing pipeline for integrated digital PET/MR detectors

In this paper, we propose an FPGA-based Data Acquisition and Control Architecture (DACA) with distributed processing capabilities for future PET-systems suitable for simultaneous PET-MRI operation, which can be configured to future detector geometries and crystal configurations of any preclinical and clinical PET system. Key-aspects for such an architecture are a scalable and modular infrastructure with common communication blocks for every processing node, a common interface for processing pipelines, bypass of raw detector data and a system-wide message-based data routing with modular addressing scheme. We present general conceptual aspects and implementations for the preclinical PET-Insert Hyperion-I as proposed by Schulz et at. and Hyperion-UD as proposed by Weissler et al., followed by initial throughput measurements of a PET-module, which demonstrates a linear response behavior of our FPGA-based architecture within the aimed data rates of up to 930Mbits/s.