25 July 2023 in IEEE Internet of Things Journal
Simple, accurate, portable, and selective gas sensors with autonomous, remote, and real-time access have become a requisite in various fields of applications. In this article, we report the development of a stand-alone and selective gas sensor system incorporating a single resistive sensor with wireless monitoring and Internet connectivity. The sensor is fabricated in-house with platinum-decorated tin-oxide hollow spheres as the sensing material, which exhibits a prominent response toward the tested volatile organic compounds (VOCs) at different concentrations. The intelligence in terms of accurate identification of VOCs and their concentration is attained by employing a machine learning tool based on a deep neural network. The applied model displays an average accuracy of 96.43% with a fast prediction speed of $310~\mu \text{s}$ , allowing a real-time recognition capability. The wireless connectivity is established utilizing a low-power microcontroller board and a Bluetooth module. The real-time data is made available for the users over an Android-based mobile application and a webpage while utilizing cloud services through the Internet. The implemented system is successfully experimented with and validated under different test conditions that verify the whole platform. Further, the sensor system can be potentially applied to a remote application without needing any manual involvement. The demonstrated work with an Internet of Things (IoT) paradigm strengthens the next-generation gas sensing technology for developing smart, selective, and real-time gas sensor systems.
30 November 2020 in IEEE Sensors Journal
The paper presents the integration of single metal-oxide based chemiresistive sensor device and machine learning tools for selective discrimination of different volatile organic compounds (VOCs) for indoor air quality monitoring applications. Tungsten oxide (WO 3 ) nanoplates has been employed as the gas sensing material which were obtained by acidification followed by low temperature hydrothermal process. Synthesized WO 3 nanoplate structure was confirmed by different characterization tools explaining surface morphology and structural properties. The sensor device was fabricated by using a simple drop coating technique on top of aluminum based interdigitated electrodes. An extensive gas sensing study was carried out where adequate sensor response was observed for each target VOC. The sensing mechanism has been discussed to realize the behavior of the sensor towards the introduction of target VOCs. Collective data obtained from the sensor device were engaged with machine learning algorithms (best results shown by multilayer perceptron) to discriminate the target VOCs accurately. Furthermore, concentrations of tested VOCs were predicted in a quantitative manner using a regression model with fair accuracy.
25 November 2019 in IEEE Transactions on Circuits and Systems II: Express Briefs
This brief proposes and demonstrates a reliable power and data transfer scheme for biomedical implants through a resonant capacitive link. Flexible, conformable, and biocompatible patches are adopted with an automatic resonant frequency calibration technique for the first time. A voltage feedback network, a Class-D driver, and a custom algorithm have been developed for the low overhead calibration. An ex vivo chicken skin tissue-based experiments with 4 cm2 flexible, conformal capacitive patches demonstrated the method. Experimental results show that the power delivered to the load (PDL) of the link extends up to 150 mW and maximum power transfer efficiency (PTE) of the link is 54%. A simultaneous data transfer with a rate of up to 170 kbps alongside power transfer using 1.5 to 2 mm-thick biological skin tissue has been achieved.
13 November 2023 in IEEE Transactions on Biomedical Circuits and Systems
This paper presents an innovative, minimally invasive, battery-free, wireless, peripheral nervous system (PNS) neural interface, which seamlessly integrates a millimeter-scale, fascicle-selective integrated circuit (IC) with extraneural recording and stimulating channels. The system also incorporates a wearable interrogator equipped with integrated machine-learning capabilities. This PNS interface is specifically tailored for adaptive neuromodulation therapy, targeting individuals with paralysis, amputation, or chronic medical conditions. By employing a neural pathway classifier and temporal interference stimulation, the proposed interface achieves precise deep fascicle selectivity for recording and stimulation without the need for nerve penetration or compression. Ultrasonic energy harvesters facilitate wireless power harvesting and data reception, enhancing the usability of the system. Key circuit performance metrics encompass a 2.2 μVrms input-referred noise, 14-bit ENOB, and a 173 dB Schreier figure of merit (FOM) for the neural analog-to-digital converter (ADC). Additionally, the ultra-low-power radio-frequency (RF) transmitter boasts a remarkable 1.38 pJ/bit energy efficiency. In vivo experiments conducted on rat sciatic nerves provide compelling evidence of the interface's ability to selectively stimulate and record neural fascicles. The proposed PNS neural interface offers alternative treatment options for diagnosing and treating neurological disorders, as well as restoring or repairing neural functions, improving the quality of life for patients with neurological and sensory deficits.
1 June 2021 in Microelectronics Journal