CAMEL: An Ultra-Low-Power VGA CMOS Imager based on a Time-Based DPS Array

Adding vision capabilities to wireless sensors nodes (WSN) for the Internet-of-Things requires imagers with high resolution and ultra-low power. Based on previous work [1], a time-based imager adapts the three key innovations from [3] and introduces new features: extended resolution, migration to a more mature CMOS technology and integration of a DC/DC converter on chip to power the imager through a single supply. The concepts are demonstrated by the design and fabrication of a VGA CMOS image sensor in 0.18μm CMOS. When operated at 0.65 V, it features a fixed pattern noise (FPN) of 0.078%, a dynamic range (DR) of 61.6 dB and a fill factor (FF) of 32.4% with a 5.8 μm pixel pitch, while consuming only 177.3 pJ/(frame.pixel) and 470 μW at 8.5 fps.