详细信息
双重建模与分离辨识技术的FPGA实现
Realization of Dual-Modeling and Separate-Identification Technology in FPGA Verilong-Language
文献类型:期刊文献
中文题名:双重建模与分离辨识技术的FPGA实现
英文题名:Realization of Dual-Modeling and Separate-Identification Technology in FPGA Verilong-Language
作者:曹辉[1];韩建国[2];马栋萍[1];王暄[1]
第一作者:曹辉
机构:[1]北京联合大学生物化学工程学院;[2]北京化工大学信息科学与技术学院
第一机构:北京联合大学生物化学工程学院
年份:2010
卷号:17
期号:3
起止页码:297-299
中文期刊名:控制工程
外文期刊名:Control Engineering of China
收录:CSTPCD;;北大核心:【北大核心2008】;CSCD:【CSCD2011_2012】;
基金:北京市教委资助项目(NO.KM200711417014)
语种:中文
中文关键词:非线性与多变量系统;双重建模与分离辨识;FPGA;Verilog-语言
外文关键词:non-linear and multivariable system; dual-modeling and separate-identification FPGA Verilong-language
摘要:针对许多易受强随机干扰而难以用常规方法辨识的动态系统,提出一种非线性多变量动态系统的双重建模与分离辨识技术在FPGA上的实现。通过对非线性子通道特性分析,采用DDSI技术进行建模,包括构造BP神经网络、运用最小二乘辨识等算法获取数学模型,其中,建立模型的主要算法和系统参数估计的相关的FPGA结构设计通过Verilog语言的结构描述数据流进行了描述。同时采用FPGA的在线可重构技术,在运行时根据需要动态改变系统的电路结构,使硬件具有分时复用,节省逻辑资源的优良性能。这一技术的FPGA实现,使得这种非线性、多变量动态系统的特殊系统辨识技术能够广泛应用于诸如星体运动、现代控制过程、生产过程与经济管理系统等高层次科技领域中的带有强噪声干扰子通道的系统动态建模中。
The FPGA-realization of dual-modeling and separate-identification of non-linear and multivariable dynamical systems is introduced to the problem that many systems with white-noise-disturbances are impossible to be modeled by traditional method.By analyzing the characteristic of nonlinear sub-channels,the system model is obtained by using the DDSI technology,including BP neural network and the method of least squares,etc.The corresponding FPGA structure design for the major algorithm and system parameter estimation is described by using Verilog structure description data stream.By the dynamic recofigurable FPGA,the configration of the circuit can be changed as need.And the hardware has excellent performance of TDMA and economizes logic source.The important special solution technology of nonlinear multivariable dynamical system can be widely applied into some high-level science and technology fields,such as star moving,modern control process,production and economic management system,where some sub-channels of the system have strong white noise disturbances.
参考文献:
正在载入数据...