A Hybrid ADC Based on Hopfield Neural Network SAR Architecture with First-Order Passive Noise Shaping

カテゴリ：研究会(論文単位)

論文No：ECT25050

グループ名：【C】電子・情報・システム部門 電子回路研究会

発行日：2025/9/1

タイトル(英語)：A Hybrid ADC Based on Hopfield Neural Network SAR Architecture with First-Order Passive Noise Shaping

著者名：PAN QINGFENG(Tokyo City University),SAN HAO(Tokyo City University)

著者名(英語)： QINGFENG PAN(Tokyo City University),HAO SAN(Tokyo City University)

キーワード：Hopfield Neural-Network ADC,SAR(Successive Approximation) ADC ,Passive Noise-Shaping,Hybrid Mixed-Signal Architecture,High-Accuracy and High-Speed ADC,Low-Power IoT Sensor Interface

要約(日本語)：This work proposes a hybrid SAR ADC that combines a Hopfield analog neural-network core with a first-order passive noise-shaping residue integrator, implemented using only capacitors and switches, to meet the low-power, high-accuracy, and high-speed requirements of IoT sensors. The design improves accuracy over conventional Hopfield-SAR converters while adding negligible power, making it well-suited for energy-constrained IoT applications.

要約(英語)：This work proposes a hybrid SAR ADC that combines a Hopfield analog neural-network core with a first-order passive noise-shaping residue integrator, implemented using only capacitors and switches, to meet the low-power, high-accuracy, and high-speed requirements of IoT sensors. The design improves accuracy over conventional Hopfield-SAR converters while adding negligible power, making it well-suited for energy-constrained IoT applications.

本誌：2025年9月4日電子回路研究会

本誌掲載ページ：45-46p

原稿種別：英語

PDFファイルサイズ：554Kバイト