← 返回 JSSC 论文列表JSSC 2024第1期Memory40nmSRAMEmerging Memory
A Heterogeneous RRAM In-Memory and SRAM Near-Memory SoC for Fused Frame and Even
提出一种异构RRAM内存计算和SRAM近内存计算的SoC,用于融合帧和事件视觉处理。
40nm TSMC ULP工艺
RRAM内存计算SRAM近内存计算异构SoC帧和事件融合电源门控
▸异构RRAM内存计算和SRAM近内存计算的SoC设计
▸两级别电源门控技术节省91.8%总功耗
▸嵌入式三重错误校正(TEC)技术
Abstract
Accurate identification of the target and tracking
it at high speeds using drone-mounted cameras and compute
hardware finds military and commercial applications. Conven-
tional frame-based cameras and convolutional neural networks
(CNNs) extract detailed spatial information to show high accu-
racy but suffer from lower throughput caused by large models.
Alternatively, event cameras capture the motion information as
an asynchronous event stream with high temporal resolution.
Spiking neural networ