Hand sign recognition is gaining importance in human-human and in human-machine communication and interaction. Electrical Impedance Tomography (EIT) is thereby very interesting as it provides information on impedance changes in depth of the Section of the arm, which infers muscle contractions. This paper introduces an EIT imaging system for hand sign recognition and monitoring having a low complexity and including an electronic interface with 8 electrodes placed on the forearm, a Gauss-Newton image reconstruction algorithm, a robust CNN based hand sign classification and a virtual hand model for visualization. A database has been collected for EIT measurements in pole mode taken by eight subjects performing the American sign language numbers from 0 to 9. The overall imaging system is validated using a water tank system, where conductive objects can be changed in properties and positions. The correspondence between the reconstructed images and the expected muscle behavior for the hand signs is investigated. A robust Convolutional Neural Network (CNN) classification algorithm was implemented and optimized by implementing an Adam optimizer and conducting a dedicated study to avoid overfitting. The results obtained by CNN are compared to the results by a Support Vector Machine (SVM), and a Softmax classifier. They show a classification accuracy of 95.94%, 75.61%, and 62.9% respectively. In term of subject dependency, the system using the CNN model shows a higher performance, as the accuracy decreases only by 0.72% while increasing the number of subjects from one to eight. Finally, for visualization, a 3D virtual hand model is designed and controlled based on detected hand signs.

We investigated functionalized graphene materials to create highly sensitive sensors for volatile organic compounds (VOCs) such as formaldehyde, methanol, ethanol, acetone, and isopropanol. First, we prepared VOC-sensitive films consisting of mechanically exfoliated graphene (eG) and chemical graphene oxide (GO), which have different concentrations of structural defects. We deposited the films on silver interdigitated electrodes on Kapton substrate and submitted them to thermal treatment. Next, we measured the sensitive properties of the resulting sensors towards specific VOCs by impedance spectroscopy. We obtained the eG- and GO-based electronic nose composed of two eG films- and four GO film-based sensors with variable sensitivity to individual VOCs. The smallest relative change in impedance was 5% for the sensor based on eG film annealed at 180 °C toward 10 ppm formaldehyde, whereas the highest relative change was 257% for the sensor based on two-layers deposited GO film annealed at 200 °C toward 80 ppm ethanol. At 10 ppm VOC, the GO film-based sensors were sensitive enough to distinguish between individual VOCs, which implied excellent selectivity, as confirmed by Principle Component Analysis (PCA). According to a PCA-Support Vector Machine-based signal processing method, the electronic nose provided identification accuracy of 100% for individual VOCs. The proposed electronic nose can be used to detect multiple VOCs selectively because each sensor is sensitive to VOCs and has significant cross-selectivity to others.

For embedded impedance spectroscopy, a suitable method for analyzing AC signals needs to be carefully chosen to overcome limited processing capability and memory availability. This paper compares various methods, including the fast Fourier transform (FFT), the FFT with barycenter correction, the FFT with windowing, the Goertzel filter, the discrete-time Fourier transform (DTFT), and sine fitting using linear or nonlinear least squares, and cross-correlation, for analyzing AC signals in terms of speed, memory requirements, amplitude measurement accuracy, and phase measurement accuracy. These methods are implemented in reference systems with and without hardware acceleration for validation. The investigation results show that the Goertzel algorithm has the best overall performance when hardware acceleration is excluded or in the case of memory constraints. In implementations with hardware acceleration, the FFT with barycentre correction stands out. The linear sine fitting method provides the most accurate amplitude and phase determinations at the expense of speed and memory requirements.

Zero potential circuit (ZPC) is widely implemented for two-dimensional resistive sensor array/matrix measurement to eliminate the crosstalk effect caused by the additional current paths through the un-targeted sensors. However, the measurement accuracy of the ZPC method is strongly influenced by its applied multiplexer switch-on resistors and operational amplifier DC offset voltage. This paper proposes to supply simultaneously each row of the sensor matrix with an AC voltage at a row-specific frequency, so that the responses of sensors in one column can be isolated from each other based on a frequency spectrum analysis and all sensors can be therefore scanned simultaneously. Comparing to the original ZPC method, this new approach does not need the un-targeted sensors to be short-circuited, as well as the multiplexer can be omitted. The resistance value of each sensor is calculated by evaluating its column output signal at the frequency of its corresponding row excitation AC voltage source. The DC offset voltage of the operational amplifier does not influence the measurement anymore because of the AC measurement principle.

Impedance measurement ICs are widely used in embedded systems for single-sensor measurements. The impedance value of the target sensor is calculated by comparing its measurement signal with one of a given reference element. This is known as the calibration concept. To automate the calibration process and to enable the multi-sensor measurements, multiplexers are recommended. But the switch on-resistances of multiplexers cause additional measurement deviations, unless expensive multiplexers with very low switch on-resistances or buffer functions are used. To overcome the need for specially designed multiplexers, a circuit structure based on two multiplexers in the feedback loop of one voltage follower is proposed in this work, with DC bias circuits. This solution is experimentally validated in a prototype system based on the AD5933 chip for a one-dimensional sensor matrix. Compared to the common single multiplexer system, the proposed approach has suppressed the average measurement deviation from 6.01 % to 0.06 % in the impedance magnitude measurements, and from 10.75 % to 0.13 % for the resistive targets.

The alternating current zero potential circuit (AC-ZPC) has been recently proposed for two-dimensional resistive sensor matrices, where each matrix row is driven by an AC signal at an individual frequency and every column is read separately. The AC-ZPC method enables the simultaneous measurement of all sensors in the matrix and is insensitive to DC offset voltages of amplifiers. This work aims to investigate the feasibility of the AC-ZPC method for matrices of impedimetric sensors and to propose a calibration approach for suppressing the measurement deviations caused by row interface impedances. The feasibility of the AC-ZPC measurement method for the impedimetric sensor matrices and the proposed calibration approach have been proved by both simulation and implementation results. Based on the commercial impedance measurement circuit AD5933, a prototype has been realized. It achieves a measurement deviation below 1% in the range from 1.9 kΩ||361.7 pF to 15.9 k Ω||102.0 pF.