RIA: High Speed High Precision Signal Processing with Parallel Analog Circuits

本课题研究利用并行模拟电路开发高速、高精度的信号处理电路。模拟电路实现高速信号处理的能力是众所周知的。然而，高速通常是以低精度为代价来实现的，这使得高速模拟电路对于滤波和仿真等许多重要的信号处理应用来说不切实际。高速模拟电路的低精度是由于无法控制构成电路的器件（如晶体管和电容器）的物理特性造成的。在这项工作中，研究了独立于所使用的器件物理特性而实现所需功能的并行电路。这些电路设计用于执行各种各样的信号处理功能，并且在制造后通过高速训练使其非常精确。许多使用CMOS器件的并行模拟设计被用来演示该技术。

This research deals with the development of high speed, high precision signal processing circuits using parallel analog circuitry. The ability of analog circuits to achieve high speed signal processing is well known. However, high speed has generally been achieved at the expense of low precision, making high speed analog circuits impractical for many important signal processing applications such as filtering and simulation. The low precision of high speed analog circuits results from the inability to control the physical properties of the devices, such as transistors and capacitors, that the circuits are constructed from. In this work, parallel circuits that achieve the required functionality independent of the device physics used are researched. These circuits are designed to perform a large variety of signal processing functions and are made very precise by high speed training after fabrication. Many parallel analog designs using CMOS devices are made to demonstrate the technique.