摘要
在清华大学微纳器件和系统实验室制备了一种高性能的巨磁电阻(GMR)自旋阀。因为该GMR磁传感器的输出呈高度非线性,不利于后端客户的应用,所以设计了一种用于线性校正用途的模糊神经网络(FNN),并以此构建了智能GMR磁传感器系统,通过Matlab仿真试验验证了该方法的有效性。最后,讨论了单芯片系统(SOC)实现该智能GMR磁传感器的可行性,为进一步的系统集成提供了理论基础。
The high performance giant magnetoresistance （GMR） spin valves are fabricated in Tsinghua university micro/nano devices and systems lab. Since the response characteristics of this GMR sensor are highly nonlinear, it is not convenience for consumer application. A fuzzy neural network （FNN） based on linear calibration cell is proposed and a smart GMR sensor is constructed. The Matlab simulations show the effective results of this method. A system on chip （SOC） implementation scheme for this FNN-based smart GMR sensor is suggested, which is a theory fundamental for the system integration.
出处
《计算机工程与设计》
CSCD
北大核心
2007年第4期895-897,共3页
Computer Engineering and Design
基金
国家自然科学基金项目(90407013)
关键词
巨磁电阻
模糊神经网络
单芯片系统
智能传感器
线性校正
giant magnetoresistance
fuzzy neural network
system on chip
smart sensor
linear calibration
