Poor molecular selectivity and unacceptable rate of sensor-to-sensor variations have been stumbling blocks in translating chemical and biological sensors based of microelectromechanical systems into a practical reality. Traditionally, selectivity in chemical and biological sensing is achieved by using immobilized receptors or chemical interfaces on cantilever surfaces. While poor selectivity in small molecule detection using reversible receptors is due to the lack of uniqueness of receptor-analyte interaction, nonuniformity in the immobilized receptor graft density leads to unacceptable rates of reproducibility in these sensors. Engineering the receptor layers with vectorial connectivity along the length of the cantilevers shows increased reproducibility and much enhanced sensitivity in chemical and biological detection. Various strategies are currently used for overcoming the selectivity challenges. These approaches include separation prior to detection, array-based sensing, and incorporation of multi-modal, multi-physics approaches. These cantilevers, when fabricated as bimaterial cantilevers, can be used as highly sensitive thermal sensors for photothermal spectroscopy of adsorbed molecules. Cantilever-based photothermal spectroscopy overcomes many of the selectivity challenges encountered when using receptor-based approaches. Multi-modal, multi-physics data obtained by judicial integration of these approaches when analyzed using advanced machine learning techniques enhance the selectivity, sensitivity, and reliability of MEMS sensor arrays.

Continuous and robust monitoring of physiological signals with wearable devices provides new opportunities for improving the care and outcomes for people with or at risk of a broad range of adverse health events. One challenge has been that many devices are obtrusive, cumbersome, not very convenient to wear, and unsuitable for ambulatory care. Limited reliability and robustness of these sensing paradigms have also been among contributing factors that prohibit the adoption of wearable devices into the FDA-regulatory space and medical applications. Another challenge to the current method of measuring certain hemodynamic parameters such as blood pressure (BP), is that cuff-based sensors are only capable of providing infrequent measurements and are uncomfortable. Elevated BP is a critically important risk factor for various cardiovascular disorders (i.e., heart attack, stroke, heart failure), kidney diseases, vision loss, and sexual dysfunction. In this talk, we discuss our techniques that directly address these unmet needs for a device that can unobtrusively, accurately, and continuously measure BP. We seek to develop and test a novel transformative solution which addresses concerns of wearability and robust sensing, and enables new sensing paradigms that can be deployed for field-based, mobile, or ambulatory care leveraging bio-impedance. We will discuss a number of sensing, circuit and signal processing paradigms that capture physiological observations including bio-impedance. Our primary focus remains estimating blood pressure from a wrist-worn device with a watch form factor with high degrees of precision. We will discuss several methodologies for noise rejection that improve the robustness of signal acquisition. We will also discuss machine learning techniques that converts our bio-impedance observations to blood pressure. We will offer concluding remarks on the trends of wearable computing technology development and potential future directions.

Sensor placement in wireless sensor networks (WSN) aims to maximize coverage while minimizing total deployment cost. However, existing coverage-only approaches do not consider the robustness of the entire system where sensors may break down or malfunction. In this talk, we first show a robustness-aware sensor placement approach by constructing a multi-objective optimization model. We demonstrate that this method increases the robustness of a WSN by up to 50%, with 201% higher probability of monitoring the entire environment as compared to the state-of-the-art coverage-only approach. We further improve our proposed method by introducing a robust optimization-based sensor placement approach that considers the uncertainty of the distance between a sensor and a target. We show that this improved model increases the probability of target detection by up to 77% compared to state-of-the-art coverage-only approach.