基于时间数字直接转换的超低功耗CMOS图像传感器关键技术研究

The ultra low power CMOS image sensors have many important applications in the internet of things due to the advantages such as long life time, small size, low cost and so on. In the conventional CMOS image sensors, the capture and quantization of image signals are all based on the operation of the integration voltages. But the integration voltages directly affect the key performance such as the signal-to-noise ratio, dynamic range and so on, which limit the reduction of the supply voltage of the CMOS image sensor. What is more, plenty of power hungry modules such as source followers and A/D converters are used in the chip. All of these make the order of magnitude of the power consumption of the current commercial CMOS image sensor is about 100mW. Therefore, this project mainly studies the theory and key techniques of the 10μW ultra low power CMOS image sensor. In the image capture module, we study the theory of the time-digital direct conversion technique. The light signals are converted to the width of the pulse rather than the voltage in the conventional CMOS image sensor to reduce the supply voltage and power consumption greatly. In the performance enhancement module, we study the theory of the thresthold voltage cancellation technique and dynamic thresthold current technique to reduce the fixed pattern noise and increase the dynamic range. The design theory and technique mentioned above will be obtained through the theory analysis, computer simulation and chip test. The theory basis for the ultra low power CMOS image sensor will be accumulated through the studies in this project.

由于超低功耗CMOS图像传感器具有待机时间长、体积小、成本低等优点，在物联网中具有重要的实用价值。传统的CMOS图像传感器，图像信号的采集量化均基于对积分电压的操作，而积分电压的大小直接决定了信噪比和动态范围等关键性能指标的好坏，该因素严格限制了芯片电源电压的降低，加上源级跟随器和模数转换器等大电流模块的使用，导致目前商用芯片的功耗在100mW量级。为此，本项目拟针对10μW量级的超低功耗CMOS图像传感器芯片核心技术及理论进行探索研究。在图像采集模块中，拟探索时间数字直接转换技术及理论，将光强信号转换为脉冲宽度而非传统的电压信号，以大幅降低电路的工作电压和功耗；在性能增强模块中，拟研究阈值电压消除技术和动态门限电流技术及理论，以降低电路的固定模式噪声并提高动态范围。项目拟通过理论分析、仿真和实测等手段摸索出上述关键技术的设计理论和设计方法，为实现超低功耗CMOS图像传感器奠定理论基础。

由于超低功耗CMOS图像传感器具有待机时间长、体积小、成本低等优点，在物联网中具有重要的实用价值。传统的CMOS图像传感器，图像信号的采集量化均基于对积分电压的操作，而积分电压的大小直接决定了信噪比和动态范围等关键性能指标的好坏，该因素严格限制了芯片电源电压的降低，加上源级跟随器和模数转换器等大电流模块的使用，导致目前商用芯片的功耗在100mW量级。为此，本项目拟针对10μW量级的超低功耗CMOS图像传感器芯片核心技术及理论进行探索研究。在图像采集模块中，拟探索时间数字直接转换技术及理论，将光强信号转换为脉冲宽度而非传统的电压信号，以大幅降低电路的工作电压和功耗；在性能增强模块中，拟研究阈值电压消除技术和动态门限电流技术及理论，以降低电路的固定模式噪声并提高动态范围。项目拟通过芯片设计、仿真和实测等手段摸索出上述关键技术的设计理论和设计方法，为实现超低功耗CMOS图像传感器奠定理论基础。