The development of intelligent miniaturized biochemical sensors has been an area of active research over the past several years. These microsensors and sensor microarrays are finding niche applications in point-of-care diagnostics, personal care, food safety, and environmental monitoring. Among these sensors, optical (luminescence) sensing holds a great promise towards implementing simple, specific, and highly sensitive biochemical sensors. It is generally understood that biochemical recognition elements that respond specifically to the target analytes play a critical role in the overall sensor operation. Aside from the recognition elements, signal detection and processing components are important to collect the information provided by recognition elements and output an easily understandable response. The signal processing component provides the best opportunity to incorporate intelligence to achieve low-power, adaptive, accurate, and reliable sensors. We deal with sensors that use sol-gel derived xerogels as recognition materials and Complementary Metal-Oxide Semiconductor (CMOS) integrated circuits for signal detection and processing. Xerogels are nano/microporous glasses that can be used to encapsulate luminophores, enzymes, and nanoparticles in their pores. In this Article, we will describe some of the emerging integrated sensor platforms that are based on monitoring the excitedstate luminescence intensity and lifetimes of the luminophores housed in the xerogels. Specifically, we describe a CMOS imaging system for simultaneously monitoring xerogels sensor arrays. Next, we describe a non-linear phase luminometric system with enhanced and dynamically tunable sensitivity and improved signal-to-noise performance. Finally, we will describe time-based signal processing that could enable the direct measurement of excited state fluorescence lifetimes. This time-to-digital converter requires simple circuit implementation and can be used to measure lifetimes that are on the order of several hundred nanoseconds. The time based signal processing could ultimately allow the development of low-cost lifetime imaging system wherein one could take the lifetimes' image of an array of recognition elements rather than collecting an image of their fluorescence intensities.