粒子跟踪探测器中信号处理与 AI/ML 基于数据过滤的像素内集成
仪器与探测器
2025-10-15 v2 高能物理 - 实验
核实验
摘要
我们展示了粒子跟踪探测器中信号处理与 AI/ML 基于数据过滤的像素内集成的首个物理实现。基于先前工作中适合 ASIC 实现的受物理启发的边缘 AI 算法，本工作标志着向智能硅跟踪器迈出的重要里程碑。我们的原型读出芯片在传感器级别执行实时数据降低，同时满足功耗、面积和延迟等严格要求。该芯片采用 28nm TSMC CMOS Bulk 工艺制造，已证明其对粒子实验具有足够的辐射硬度。这一发展代表了实现完全在探测器上端点 AI 的关键一步，为高速率、高辐射环境下的数据吞吐量和发现潜力铺平了道路。
引用
@article{arxiv.2510.07485,
title = {In-pixel integration of signal processing and AI/ML based data filtering for particle tracking detectors},
author = {Benjamin Parpillon and Anthony Badea and Danush Shekar and Christian Gingu and Giuseppe Di Guglielmo and Tom Deline and Adam Quinn and Michele Ronchi and Benjamin Weiss and Jennet Dickinson and Jieun Yoo and Corrinne Mills and Daniel Abadjiev and Aidan Nicholas and Eliza Howard and Carissa Kumar and Eric You and Mira Littmann and Karri DiPetrillo and Arghya Ranjan Das and Mia Liu and David Jiang and Mark S. Neubauer and Morris Swartz and Petar Maksimovic and Alice Bean and Ricardo Silvestre and Jannicke Pearkes and Keith Ulmer and Nick Manganelli and Chinar Syal and Doug Berry and Nhan Tran and Lindsey Gray and Farah Fahim},
journal= {arXiv preprint arXiv:2510.07485},
year = {2025}
}