Micromachined Thermal Gas Sensors-A Review.

In recent years, there has been a growing desire to monitor and control harmful substances arising from industrial processes that impact upon our health and quality of life. This has led to a large market demand for gas sensors, which are commonly based on sensors that rely upon a chemical reaction with the target analyte. In contrast, thermal conductivity detectors are physical sensors that detect gases through a change in their thermal conductivity. Thermal conductivity gas sensors offer several advantages over their chemical (reactive) counterparts that include higher reproducibility, better stability, lower cost, lower power consumption, simpler construction, faster response time, longer lifetime, wide dynamic range, and smaller footprint. It is for these reasons, despite a poor selectivity, that they are gaining renewed interest after recent developments in MEMS-based silicon sensors allowing CMOS integration and smart application within the emerging Internet of Things (IoT). This timely review focuses on the state-of-the-art in thermal conductivity sensors; it contains a general introduction, theory of operation, interface electronics, use in commercial applications, and recent research developments. In addition, both steady-state and transient methods of operation are discussed with their relative advantages and disadvantages presented. Finally, some of recent innovations in thermal conductivity gas sensors are explored.

Optimising Deep Learning at the Edge for Accurate Hourly Air Quality Prediction.

Accurate air quality monitoring requires processing of multi-dimensional, multi-location sensor data, which has previously been considered in centralised machine learning models. These are often unsuitable for resource-constrained edge devices. In this article, we address this challenge by: (1) designing a novel hybrid deep learning model for hourly PM2.5 pollutant prediction; (2) optimising the obtained model for edge devices; and (3) examining model performance running on the edge devices in terms of both accuracy and latency. The hybrid deep learning model in this work comprises a 1D Convolutional Neural Network (CNN) and a Long Short-Term Memory (LSTM) to predict hourly PM2.5 concentration. The results show that our proposed model outperforms other deep learning models, evaluated by calculating RMSE and MAE errors. The proposed model was optimised for edge devices, the Raspberry Pi 3 Model B+ (RPi3B+) and Raspberry Pi 4 Model B (RPi4B). This optimised model reduced file size to a quarter of the original, with further size reduction achieved by implementing different post-training quantisation. In total, 8272 hourly samples were continuously fed to the edge device, with the RPi4B executing the model twice as fast as the RPi3B+ in all quantisation modes. Full-integer quantisation produced the lowest execution time, with latencies of 2.19 s and 4.73 s for RPi4B and RPi3B+, respectively.

Real-Time Thermal Modulation of High Bandwidth MOX Gas Sensors for Mobile Robot Applications.

A new signal processing technique has been developed for resistive metal oxide (MOX) gas sensors to enable high-bandwidth measurements and enhanced selectivity at PPM levels (<5 PPM VOCs). An embedded micro-heater is thermally pulsed from a temperature of 225 to 350 °C, which enables the chemical reaction kinetics of the sensing film to be extracted using a fast Fourier transform. Signal processing is performed in real-time using a low-cost microcontroller integrated into a sensor module. Three sensors, coated with SnO2, WO3 and NiO respectively, were operated and processed at the same time. This approach enables the removal of long-term baseline drift and is more resilient to changes in ambient temperature. It also greatly reduced the measurement time from ~10 s to 2 s or less. Bench-top experimental results are presented for 0 to 200 ppm of acetone, and 0 ppm to 500 ppm of ethanol. Our results demonstrate our sensor system can be used on a mobile robot for real-time gas sensing.

Ratiometric Decoding of Pheromones for a Biomimetic Infochemical Communication System.

Biosynthetic infochemical communication is an emerging scientific field employing molecular compounds for information transmission, labelling, and biochemical interfacing; having potential application in diverse areas ranging from pest management to group coordination of swarming robots. Our communication system comprises a chemoemitter module that encodes information by producing volatile pheromone components and a chemoreceiver module that decodes the transmitted ratiometric information via polymer-coated piezoelectric Surface Acoustic Wave Resonator (SAWR) sensors. The inspiration for such a system is based on the pheromone-based communication between insects. Ten features are extracted from the SAWR sensor response and analysed using multi-variate classification techniques, i.e., Linear Discriminant Analysis (LDA), Probabilistic Neural Network (PNN), and Multilayer Perception Neural Network (MLPNN) methods, and an optimal feature subset is identified. A combination of steady state and transient features of the sensor signals showed superior performances with LDA and MLPNN. Although MLPNN gave excellent results reaching 100% recognition rate at 400 s, over all time stations PNN gave the best performance based on an expanded data-set with adjacent neighbours. In this case, 100% of the pheromone mixtures were successfully identified just 200 s after they were first injected into the wind tunnel. We believe that this approach can be used for future chemical communication employing simple mixtures of airborne molecules.

Electronic Noses for Well-Being: Breath Analysis and Energy Expenditure.

The wealth of information concealed in a single human breath has been of interest for many years, promising not only disease detection, but also the monitoring of our general well-being. Recent developments in the fields of nano-sensor arrays and MEMS have enabled once bulky artificial olfactory sensor systems, or so-called "electronic noses", to become smaller, lower power and portable devices. At the same time, wearable health monitoring devices are now available, although reliable breath sensing equipment is somewhat missing from the market of physical, rather than chemical sensor gadgets. In this article, we report on the unprecedented rise in healthcare problems caused by an increasingly overweight population. We first review recently-developed electronic noses for the detection of diseases by the analysis of basic volatile organic compounds (VOCs). Then, we discuss the primary cause of obesity from over eating and the high calorific content of food. We present the need to measure our individual energy expenditure from our exhaled breath. Finally, we consider the future for handheld or wearable devices to measure energy expenditure; and the potential of these devices to revolutionize healthcare, both at home and in hospitals.

Post-CMOS wafer level growth of carbon nanotubes for low-cost microsensors--a proof of concept.

Here we demonstrate a novel technique to grow carbon nanotubes (CNTs) on addressable localized areas, at wafer level, on a fully processed CMOS substrate. The CNTs were grown using tungsten micro-heaters (local growth technique) at elevated temperature on wafer scale by connecting adjacent micro-heaters through metal tracks in the scribe lane. The electrical and optical characterization show that the CNTs are identical and reproducible. We believe this wafer level integration of CNTs with CMOS circuitry enables the low-cost mass production of CNT sensors, such as chemical sensors.

Design and Development of a Disposable Lab-on-a-Chip for Prostate Cancer Detection.

We have designed and fabricated a low-cost modular electrical and fluidic integration platform for microelectromechanical systems (MEMS) based biosensors, paving the way for a disposable, low-cost Lab-on-a-Chip. We demonstrate seamless integration using an additive manufacturing enabled "plug-and-play" platform that does not require permanent electronic or fluidic integration. This paper describes the fabrication steps and assembly of the method and highlights its advantages over the more traditional methods, such as `wire bonding' and `flip chip'. We also provide design guidelines for improved biosensing, taking transport and binding kinetics into consideration in the context of prostate cancer diagnosis. Our novel approach combined with the design guidelines, opens up new opportunities for low-cost disposable high-density MEMS-based lab-on-a-chips for biosensing applications.

Bacteria classification using Cyranose 320 electronic nose.

BACKGROUND: An electronic nose (e-nose), the Cyrano Sciences' Cyranose 320, comprising an array of thirty-two polymer carbon black composite sensors has been used to identify six species of bacteria responsible for eye infections when present at a range of concentrations in saline solutions. Readings were taken from the headspace of the samples by manually introducing the portable e-nose system into a sterile glass containing a fixed volume of bacteria in suspension. Gathered data were a very complex mixture of different chemical compounds. METHOD: Linear Principal Component Analysis (PCA) method was able to classify four classes of bacteria out of six classes though in reality other two classes were not better evident from PCA analysis and we got 74% classification accuracy from PCA. An innovative data clustering approach was investigated for these bacteria data by combining the 3-dimensional scatter plot, Fuzzy C Means (FCM) and Self Organizing Map (SOM) network. Using these three data clustering algorithms simultaneously better 'classification' of six eye bacteria classes were represented. Then three supervised classifiers, namely Multi Layer Perceptron (MLP), Probabilistic Neural network (PNN) and Radial basis function network (RBF), were used to classify the six bacteria classes. RESULTS: A [6 x 1] SOM network gave 96% accuracy for bacteria classification which was best accuracy. A comparative evaluation of the classifiers was conducted for this application. The best results suggest that we are able to predict six classes of bacteria with up to 98% accuracy with the application of the RBF network. CONCLUSION: This type of bacteria data analysis and feature extraction is very difficult. But we can conclude that this combined use of three nonlinear methods can solve the feature extraction problem with very complex data and enhance the performance of Cyranose 320.

Rapid processing of chemosensor transients in a neuromorphic implementation of the insect macroglomerular complex.

We present a biologically-constrained neuromorphic spiking model of the insect antennal lobe macroglomerular complex that encodes concentration ratios of chemical components existing within a blend, implemented using a set of programmable logic neuronal modeling cores. Depending upon the level of inhibition and symmetry in its inhibitory connections, the model exhibits two dynamical regimes: fixed point attractor (winner-takes-all type), and limit cycle attractor (winnerless competition type) dynamics. We show that, when driven by chemosensor input in real-time, the dynamical trajectories of the model's projection neuron population activity accurately encode the concentration ratios of binary odor mixtures in both dynamical regimes. By deploying spike timing-dependent plasticity in a subset of the synapses in the model, we demonstrate that a Hebbian-like associative learning rule is able to organize weights into a stable configuration after exposure to a randomized training set comprising a variety of input ratios. Examining the resulting local interneuron weights in the model shows that each inhibitory neuron competes to represent possible ratios across the population, forming a ratiometric representation via mutual inhibition. After training the resulting dynamical trajectories of the projection neuron population activity show amplification and better separation in their response to inputs of different ratios. Finally, we demonstrate that by using limit cycle attractor dynamics, it is possible to recover and classify blend ratio information from the early transient phases of chemosensor responses in real-time more rapidly and accurately compared to a nearest-neighbor classifier applied to the normalized chemosensor data. Our results demonstrate the potential of biologically-constrained neuromorphic spiking models in achieving rapid and efficient classification of early phase chemosensor array transients with execution times well beyond biological timescales.