Novel ROIC Architecture for Application in Resistive Bolometric Sensors

Circuit architectures for advanced ROIC designs have not significantly improved within recent years due to the overutilization of older technology nodes for these ROIC implementations. Additional circuit functionality within these ROIC systems has not expanded in recent years either with current SOA utilizing similar circuit constructs as seen in the circuit list provided. The general flow of these ROIC systems is a multiplier after the sensor that outputs to some type of amplifier design that in turn leads to a Sigma-Delta analog-to-digital converter. Current designs within older technology nodes are missing-out on performance improvements that moving to a newer technology offers, from reduced power consumption to improved data throughput, to improved control of circuit operation. Especially with the maturation and incorporation of neural network primitives in a variety of different applications, ROIC performance metrics could significantly improve. Therefore, current utilization of ROIC designs is pointedly lacking in subthreshold/low power design, incorporation of new network constructs for improved functionality, and utilization of newer technology nodes, which the combination of all these leads to the development of a novel ROIC architecture for use with bolometers. The Nanomatronix-UA team aims to develop a low power ROIC architecture that incorporates new functionality within an advanced technology node. Utilizing Global Foundry’s 22 nm SOI technology, we plan to incorporate neural network primitives with analog/mixed-signal design methodologies used to develop low power electronics within the ROIC framework that complements typical functionality within the system.