【作者】 贾治安；

【导师】 杨文晖；

【作者基本信息】 中国科学院研究生院（电工研究所） ， 电工理论与新技术， 2005， 硕士

【摘要】 射频(RF)信号收发器是MRI 谱仪的重要组成部分,它负责MR 信号的接收和RF脉冲的调制发射。本论文旨在研究全数字高性能MRI射频收发器的原理、算法和硬件实现。首先,本文从信号处理的角度系统的描述了MRI 射频信号收发器的原理和算法。提出全数字正交解调和多级滤波抽取算法相结合用于高精度MR 信号接收。通过分析对比各种脉冲波形设计和幅度调制方法,提出了Gaussian-Sinc 脉冲和SLR 算法结合解决脉冲波形设计问题。在分析双边带和单边带调制的性能后,采用移相法实现RF 脉冲的单边带幅度调制,并给出了具体的实现算法和仿真。其次,基于软件无线电提出了一个全数字新型MRI 射频收发器硬件设计方案。这个新方案以高速AD、DA 转换器为基础,采用DSP/FPGA 器件作为系统核心,因而可以获得高精度和极大的灵活性。通过对系统的要求和难点的详细分析,给出了详细的硬件系统模块图。最后,利用TMS320C6713DSK 为硬件平台,对系统中的DSP 模块进行了模拟和仿真。采用C 语言和汇编程序混合编程完成了多级滤波抽取算法的实现和优化,在整个处理过程中采用单精度浮点数据格式,提高了处理精度。C6713DSK板上的仿真结果验证了DSP 模块中算法的正确性和实时性。更多还原

【Abstract】 RF signal transceiver is an important component of MRI spectrometer. It is responsible for the receive of MR signal and the transmit of RF pulse. This dissertation focused on the research of all-digital RF signal transceiver with high performance, including principles, algorithms and hardware implementation. Firstly, the principles and algorithms of MRI signal transceiver were described and designed. An new algorithm combined all-digital quadrature demodulation and multi-stage decimation was presented in detail. By comparing and analyzing several shaped pulse design methods were compared and analyzed, and a composite method was presented for various MRI pulse sequences. After comparing with the performance of double sideband modulation and single sideband modulation, Single sideband modulation based on phasing was used for RF pulse modulation, which can cut down the peak power of RF pulse transmitter and achieve a precise phase. Secondly, a novel MRI RF transmitter hardware architecture based on software-radio idea was presented, which used DSP and FPGA devices including high speed ADC/DAC chipsets. Advantages of this solution are of high precision and good flexibility. After the system level design and concerns were analyzed, the detailed chart of hardware platform was given. Finally, DSP module simulation based on TMS320C6713 DSK was demonstrated in detail. An optimized filter-decitmation method was established and implemented with a DSP program in ANSI C and inline assembly, and Single-precision floating-point format was used in the whole process so that large dynamic range and high processing precision are achieved. The result of simulation on 6713DSK proved that this algorithm satisfied the real-time requirement.更多还原