Low-Power VLSI Systems for Signal Sensing and Processing

人们对电子系统在计算能力、移动性和自主性方面的性能要求越来越高，以扩大其在通信、生物医学工程和远程控制等各个领域的应用。功耗是VLSI系统设计和实现中的一个关键问题，因为它限制了集成电路的密度和电池供电的移动/自主系统的寿命。在一个信号检测和处理系统中涉及三种电路。它们是传感器、传感器调理电路和信号处理电路。传感器将物理变量转换为电流或电压信号，传感器调理电路用于使传感器产生的信号适应处理电路。传感器、传感器调理电路和处理电路的低功耗设计降低了系统的总功耗。拟议研究计划的总体目标是为这三种电路开发低功耗设计技术。它有三个方面：1)传感器电路中的超低能级能量采集。它的目的是设计集成传感器，能够收集微小空间中可用的低水平能量，并利用这些能量使传感器自我支持。2)开发超低功率传感器调节电路。调理电路要求具有很高的灵敏度，能够在很低的电平信号下工作，动态范围宽，结构简单，实现低功耗。3)开发高能效的信号处理算法和体系结构。提出的低功耗设计技术的目标是在标准的硅CMOS制造工艺中实现。实现技术目标的一个组成部分将是纳入和培训大量研究生，使其成为各行业所需的高素质人员。

There are increasing demands for a better performance of the electronic systems, in terms of computation capacity, mobility, and autonomy, to expand their applications in various areas such as communications, biomedical engineering, and remote controls. Power dissipation is a critical concern in the design and implementation of VLSI systems as it limits the density of integrated circuits and the life span of battery-operating mobile/autonomous systems. Three kinds of electronic circuits are involved in a system performing signal sensing and processing. They are sensors, sensor conditioning circuits, and signal processing circuits. A sensor converts a physical variable into a current or voltage signal, and a sensor conditioning circuit is used to make the signal produced by a sensor adapt to the processing circuits. The low-power design of sensors, sensor conditioning circuits, and processing circuits leads to a reduction of the total power dissipation of the system. The overall objective of the proposed research program is to develop low-power design techniques for the three kinds of circuits. It is three-fold as presented below.1) Ultra-low level energy harvesting in sensor circuits. It is to design integrated sensors capable of collecting low level energy available in a tiny space and of using it to make the sensors self-supporting.2) Developing ultra-low-power sensor conditioning circuits. The conditioning circuits should have a high sensitivity to operate with a very low level signals, a wide dynamic range, and simple structures to achieve low-power dissipation.3) Developing energy-efficient algorithms and architectures for signal processing. The proposed low-power design techniques would aim for the implementation in standard silicon CMOS fabrication technologies. Integral to achieving the technical objective will be the inclusion and training of a large number of graduate students to become highly qualified personnel needed for the industries.