Advanced design techniques for electronic nose integrated circuits

An electronic nose (enose) is a device that can detect and classify smells and flavors. Enose has received considerable interest in the past decade owing to its importance in several applications such as health care, environmental monitoring, robotics, automobile, food, and cosmetics. The availability of a smart miniaturized low-power olfactory system, similar to vision and hearing systems, will enable imparting one of the most important sense, olfaction, to mobile and battery operated devices. That will result in immense progress in point-of-care testing and contribute to the design of new generation of smartphones and mobile apps that integrate smelling feature, in addition to vision and hearing, to greatly improve the quality of life. In recent years, artificial intelligence (AI) has revolutionized the field of machine vision. Its applications, such as face recognition, object detection and autopilot, are currently driving the market in the high tech industries and have become an integral part of our daily lives. Now, it is the right time to culminate on the progress on AI and machine learning and the maturity of complementary metal oxide semiconductor (CMOS) integrated circuits to design a smart low power miniaturized enose targeting mobile and battery operated devices. An enose is mainly composed of three parts: a sensory array, a signal conditioning readout, and a signal processing system. However, there are still obstacles in developing a miniaturized system on chip (SoC) enose that can be used for a wide variety of applications due to lack of robustness in integrated circuit design and advanced signal processing in existent enose devices. The objective of this proposal is to develop novel techniques to design a miniature and energy efficient next generation enose integrated circuit. This device will leverage the power of silicon based sensor, sophisticated on chip analog, digital, and mixed signals processing, and artificial intelligence, to produce truly robust, adaptable, and scalable enose platform. Some of the design specific steps includes: i) implementing new chip designs that are compatible with depositing large numbers of chemically different sensing layers to create the analog receptors, ii) collecting a massive amount of data from the dense sensing array, iii) processing and digitizing the data, and iv) applying a real time advanced signal processing and AI/machine learning techniques to classify the different smells and determine their concentration levels.