| Sensors can be found everywhere in our daily life. These sensors play an important role in various environmental, industrial, security and defense, and medical applications. To date, people continue to build sensor systems that are more sophisticated, more efficient, more portable, and lower-cost to sense ever more complex non-electronic signals in the world. Because complimentary metal-oxide semiconductor (CMOS) technology enables low power, economical, and miniaturized sensor systems which can be easily mass-produced, these are the foundation of the work described in this dissertation. Specifically, the CMOS-based optical sensor systems are intended for applications such as chemical/biochemical analysis, machine vision and environmental monitoring, among others. The work presented in this dissertation includes four parts: (1) A self-organized mammalian retina model for artificial vision, (2) A CMOS neuromorphic optical sensor chip with color change-intensity change disambiguation (CCICD), (3) an autonomous bi-channel CMOS optical sensor system for real-time environmental monitoring, and (4) a CMOS oxygen sensor system based on fluorescence lifetime detection with high sensitivity and adaptability.First, I describe a 1-D and a 2-D self-organized mammalian retina models which mimic a monochromatic cone pathway. In this software model, the "input light" information is encoded into electrical pulse signals according to the light intensity. An important feature for these retina models is self-organization they automatically determine the appropriate number of cells which are needed to complete the model for a given light input pattern.Second, I present a novel terrestrial retina-like neuromorphic sensor chip can first, perform irradiance and color detection via two independent pathways simultaneously and second, perform color change-intensity change disambiguation (CCICD) function. The irradiance detection pathway has a wide-dynamic detection range, and is robust to background light variation. The color detection pathway can measure a wide wavelengths ranging from 400nm to 900nm and have a 22nm wavelength resolution.Based on these features, the sensor chip can be easily applied to various applications that require visible light information processing. As an example, an autonomous sensor system for real-time environmental monitoring is presented. This sensor system has been demonstrated to detect up to 1% phosphate concentration change of a liquid sample and measure pH of a water sample ranging from pH1 to pH9.Finally, I present a CMOS oxygen sensor system based on fluorescence lifetime detection with higher sensitivity and adaptability. This oxygen sensor system is based on the well-know principles of fluorescence spectroscopy and specifically on fluorescence quenching. A very attractive feature of this sensor microsystem is that it can be integrated with a variety of sensing materials to measure chemical and biological analytes' concentration. A sol-gel derived xerogel based sensing element is used to encapsulate analyze specific fluorophores in nanoscopic porous structure. The major contribution of this work compared with previous works is a highly increased sensitivity to the analytes of interest and the adaptability to fluorophores which have shorter lifetime. Both of them are achieved through an enhanced detection scheme that can operate at frequencies up to 1MHz. Finally, this sensor system has advanced resolution to detect oxygen concentration without contaminating the target analyte. |
