Energy-Efficient Motion-Triggered IoT CMOS Image Sensor With Capacitor Array-Assisted Charge-Injection SAR ADC

We propose a low-power image sensor with a motion-based triggering feature for the Internet of Things (IoT) applications. The sensor supports the near-pixel [within analog-to-digital converter (ADC)] motion-detection mode run on a heavily subsampled frame (32 $\times $ 20 pixels, while consuming only 1.7 $\mu \text{J}$ /frame) to trigger full-array capture only when a significant change has happened in the scene. Also, to maximize energy efficiency and image quality of the sensor, we propose a column-parallel capacitor array-assisted charge-injection SAR ADC that achieves 10b operation with readout noise of 226 $\mu V_{\text {rms}}$ , which can be lowered to 104 $\mu V_{\text {rms}}$ using multiple sampling. The ADC is highly energy efficient with best-in-class energy consumption of 63.6 pJ/frame/pixel and ADC Figure of Merit (FoM) of 14.4 $\mu V_{\text {rms}}\cdot $ nJ in full-array capture mode.