Computer-Vision- and Deep-Learning-Based Determination of Flow Regimes, Void Fraction, and Resistance Sensor Data in Microchannel Flow Boiling.

The aim of this article is to introduce a novel approach to identifying flow regimes and void fractions in microchannel flow boiling, which is based on binary image segmentation using digital image processing and deep learning. The proposed image processing pipeline uses adaptive thresholding, blurring, gamma correction, contour detection, and histogram comparison to separate vapor from liquid areas, while the deep learning method uses a customized version of a convolutional neural network (CNN) called U-net to extract meaningful features from video frames. Both approaches enabled the automatic detection of flow boiling conditions, such as bubbly, slug, and annular flow, as well as automatic void fraction calculation. Especially CNN demonstrated its ability to deliver fast and dependable results, presenting an appealing substitute to manual feature extraction. The U-net-based CNN was able to segment flow boiling images with a Dice score of 99.1% and classify the above flow regimes with an overall classification accuracy of 91%. In addition, the neural network was able to predict resistance sensor readings from image data and assign them to a flow state with a mean squared error (MSE) < 10-6.

Adaptivity of a leaf-inspired wind energy harvester with respect to wind speed and direction.

Environmental wind is a random phenomenon in both speed and direction, though it can be forecasted to some extent. An example of that is a gust which is an abrupt, but short-time change in wind speed and direction. Being a free and clean source for small-scale energy scavenging, attraction of wind is rapidly growing in the world of energy harvesters. In this paper, a leaf-like flapping wind energy harvester is introduced as the base structure in which a short-span airfoil is attached to the free end of a double-deck cantilever beam. A flap mechanism inspired by scales on sharks' skin and a tail mechanism inspired by birds' horizontal tail are proposed for integration to the base harvester to make it adaptive with respect to wind speed and direction, respectively. The use of the flap mechanism increases the leaf flapping frequency by +2.1 to +11.5 Hz at wind speeds of 1.5 to 6.0 m s-1. Therefore, since the output power of a vibrational harvester is a function of vibration frequency, a figure of merit or an efficiency parameter related to the output power will increase, as well. On the other hand, if there is a misalignment between the harvester's heading and wind direction due to change of the latter one, the harvesting performance deteriorates. Although the base harvester can realign in certain ranges of sideslip angle at each wind speed, when the tail mechanism is integrated into that, it broadens the range of realignable sideslip angles at all the investigated wind speeds by up to 80∘.

Soft octopus-inspired suction cups using dielectric elastomer actuators with sensing capabilities.

Bioinspired and biomimetic soft grippers are rapidly growing fields. They represent an advancement in soft robotics as they emulate the adaptability and flexibility of biological end effectors. A prominent example of a gripping mechanism found in nature is the octopus tentacle, enabling the animal to attach to rough and irregular surfaces. Inspired by the structure and morphology of the tentacles, this study introduces a novel design, fabrication, and characterization method of dielectric elastomer suction cups. To grasp objects, the developed suction cups perform out-of-plane deflections as the suction mechanism. Their attachment mechanism resembles that of their biological counterparts, as they do not require a pre-stretch over a rigid frame or any external hydraulic or pneumatic support to form and hold the dome structure of the suction cups. The realized artificial suction cups demonstrate the capability of generating a negative pressure up to 1.3 kPa in air and grasping and lifting objects with a maximum 58 g weight under an actuation voltage of 6 kV. They also have sensing capabilities to determine whether the grasping was successful without the need of lifting the objects.

Internal Rotation Measurement of the Knee with Polymer-Based Capacitive Strain Gauges versus Mechanical Rotation Measurement Taking Gender Differences into Account: A Comparative Analysis.

With the conventional mechanical rotation measurement of joints, only static measurements are possible with the patient at rest. In the future, it would be interesting to carry out dynamic rotation measurements, for example, when walking or participating in sports. Therefore, a measurement method with an elastic polymer-based capacitive measuring system was developed and validated. In our system, the measurement setup was comprised of a capacitive strain gauge made from a polymer, which was connected to a flexible printed circuit board. The electronics integrated into the printed circuit board allowed data acquisition and transmission. As the sensor strip was elongated, it caused a change in the spacing between the strain gauge's electrodes, leading to a modification in capacitance. Consequently, this alteration in capacitance enabled the measurement of strain. The measurement system was affixed to the knee by adhering the sensor to the skin in alignment with the anterolateral ligament (ALL), allowing the lower part of the sensor (made of silicone) and the circuit board to be in direct contact with the knee's surface. It is important to note that the sensor should be attached without any prior stretching. To validate the system, an in vivo test was conducted on 10 healthy volunteers. The dorsiflexion of the ankle was set at 2 Nm using a torque meter to eliminate any rotational laxity in the ankle. A strain gauge sensor was affixed to the Gerdii's tubercle along the course of the anterolateral ligament, just beneath the lateral epicondyle of the thigh. In three successive measurements, the internal rotation of the foot and, consequently, the lower leg was quantified with a 2 Nm torque. The alteration in the stretch mark's length was then compared to the measured internal rotation angle using the static measuring device. A statistically significant difference between genders emerged in the internal rotation range of the knee (p = 0.003), with female participants displaying a greater range of rotation compared to their male counterparts. The polymer-based capacitive strain gauge exhibited consistent linearity across all measurements, remaining within the sensor's initial 20% strain range. The comparison between length change and the knee's internal rotation angle revealed a positive correlation (r = 1, p < 0.01). The current study shows that elastic polymer-based capacitive strain gauges are a reliable instrument for the internal rotation measurement of the knee. This will allow dynamic measurements in the future under many different settings. In addition, significant gender differences in the internal rotation angle were seen.

Digital measurement of anterolateral knee laxity using strain sensors.

PURPOSE: The ambition of the research group was to develop a sensor-based system that allowed the transfer of results with strain sensors applied to the knee joint. This system was to be validated in comparison to the current static mechanical measurement system. For this purpose, the internal rotation laxity of the knee joint was measured, as it is relevant for anterolateral knee laxity and anterior cruciate ligament (ACL) injury. METHODS: This is a noninvasive measurement method using strain sensors which are applied to the skin in the course of the anterolateral ligament. The subjects were placed in supine position. First the left and then the right leg were clinically examined sequentially and documented by means of an examination form. 11 subjects aged 21 to 45 years, 5 women and 6 men were examined. Internal rotation of the lower leg was performed with a torque of 2 Nm at a knee flexion angle of 30°. RESULTS: Comparison of correlation between length change and internal knee rotation angle showed a strong positive correlation (r = 1, p < 0.01). Whereas females showed a significant higher laxity vs. males (p = 0.003). CONCLUSIONS: The present study showed that the capacitive strain sensors can be used for reproducible measurement of anterolateral knee laxity. In contrast to the previous static systems, a dynamic measurement will be possible by this method in the future.

MEMS Shielded Capacitive Pressure and Force Sensors with Excellent Thermal Stability and High Operating Temperature.

In this paper, we present an innovative manufacturing process for the production of capacitive pressure and force sensors with excellent thermal stability for high-temperature applications. The sensors, which are manufactured from a stack of two silicon chips mounted via with gold-silicon (Au-Si) or aluminum-silicon (Al-Si) eutectic bonding, are shielded, miniaturized, and allow an operating temperature of up to 500 °C. Compared to conventional methods, the greatest benefit of the manufacturing process is that different sensor dimensions can be produced in the same batch for a wide measuring range, from mN to kN. The characterization of the realized sensors shows a high linearity and a low temperature drift of 99.992% FS and -0.001% FS/K at 350 °C, as well as a nonlinearity of 0.035% FS and a temperature drift of -0.0027% FS/K at 500 °C.

Estimation of Leaf Area Index with a Multi-Channel Spectral Micro-Sensor for Wireless Sensing Networks.

The leaf area index (LAI) is a key parameter in the context of monitoring the development of tree crowns and plants in general. As parameters such as carbon assimilation, environmental stress on carbon, and the water fluxes within tree canopies are correlated to the leaves surface, this parameter is essential for understanding and modeling ecological processes. However, its continuous monitoring using manual state-of-the-art measurement instruments is still challenging. To address this challenge, we present an innovative sensor concept to obtain the LAI based on the cheap and easy to integrate multi-channel spectral sensor AS7341. Additionally, we present a method for processing and filtering the gathered data, which enables very high accuracy measurements with an nRMSE of only 0.098, compared to the manually-operated state-of-the-art instrument LAI-2200C (LiCor). The sensor that is embedded on a sensor node has been tested in long-term experiments, proving its suitability for continuous deployment over an entire season. It permits the estimation of both the plant area index (PAI) and leaf area index (LAI) and provides the first wireless system that obtains the LAI solely powered by solar cells. Its energy autonomy and wireless connectivity make it suitable for a massive deployment over large areas and at different levels of the tree crown. It may be upgraded to allow the parallel measurement of photosynthetic active radiation (PAR) and light quality, relevant parameters for monitoring processes within tree canopies.

Batch Fabrication of a Polydimethylsiloxane Based Stretchable Capacitive Strain Gauge Sensor for Orthopedics.

Polymer-based capacitive strain gauges are a novel and promising concept for measuring large displacements and strains in various applications. These novel sensors allow for high strain, well above the maximum values achieved with state-of-the-art strain gauges (Typ. 1%). In recent years, a lot of interest in this technology has existed in orthopedics, where the sensors have been used to measure knee laxity caused by a tear of the anterior cruciate ligament (ACL), and for other ligament injuries. The validation of this technology in the field has a very low level of maturity, as no fast, reproducible, and reliable manufacturing process which allows mass production of sensors with low cost exists. For this reason, in this paper, a new approach for the fabrication of polymer-based capacitive strain gauges is proposed, using polydimethylsiloxane (PDMS) as base material. It allows (1) the fast manufacturing of sensor batches with reproducible geometry, (2) includes a fabrication step for embedding rigid electrical contacts on the sensors, and (3) is designed to produce sensor batches in which the size, the number, and the position of the sensors can be adapted to the patient's anatomy. In the paper, the process repeatability and the robustness of the design are successfully proven. After 1000 large-strain elongation cycles, in the form of accelerated testing caused much higher strains than in the above-mentioned clinical scenario, the sensor's electrical contacts remained in place and the functionalities were unaltered. Moreover, the prototype of a patient customizable patch, embedding multiple sensors, was produced.

A Novel Approach to Obtain PAR with a Multi-Channel Spectral Microsensor, Suitable for Sensor Node Integration.

We propose a novel approach to measure photosynthetically active radiation (PAR ) in the form of photosynthetic photon flux density with an inexpensive, small multi-channel spectrometer sensor, with integrated optical filters and analog-to-digital converter. Our experiments prove that the combination of eight spectral channels with different optical sensitivities, such as the sensorchip in use (AS7341, ams), derive the PAR with an accuracy of 14/m2/s. Enabled by the sensor architecture, additional information about the light quality can be retrieved which is expressed in the RLQI. A calibration method is proposed, and exemplary measurements are performed. Moreover, the integration in a solar-powered wireless sensor node is outlined, which enables long-term field experiments with high sensor densities and may be used to obtain important indexes, such as the GPP.

A Comparison of Energy-Efficient Seizure Detectors for Implantable Neurostimulation Devices.

Introduction: About 30% of epilepsy patients are resistant to treatment with antiepileptic drugs, and only a minority of these are surgical candidates. A recent therapeutic approach is the application of electrical stimulation in the early phases of a seizure to interrupt its spread across the brain. To accomplish this, energy-efficient seizure detectors are required that are able to detect a seizure in its early stages. Methods: Three patient-specific, energy-efficient seizure detectors are proposed in this study: (i) random forest (RF); (ii) long short-term memory (LSTM) recurrent neural network (RNN); and (iii) convolutional neural network (CNN). Performance evaluation was based on EEG data (n = 40 patients) derived from a selected set of surface EEG electrodes, which mimic the electrode layout of an implantable neurostimulation system. As for the RF input, 16 features in the time- and frequency-domains were selected. Raw EEG data were used for both CNN and RNN. Energy consumption was estimated by a platform-independent model based on the number of arithmetic operations (AOs) and memory accesses (MAs). To validate the estimated energy consumption, the RNN classifier was implemented on an ultra-low-power microcontroller. Results: The RNN seizure detector achieved a slightly better level of performance, with a median area under the precision-recall curve score of 0.49, compared to 0.47 for CNN and 0.46 for RF. In terms of energy consumption, RF was the most efficient algorithm, with a total of 67k AOs and 67k MAs per classification. This was followed by CNN (488k AOs and 963k MAs) and RNN (772k AOs and 978k MAs), whereby MAs contributed more to total energy consumption. Measurements derived from the hardware implementation of the RNN algorithm demonstrated a significant correlation between estimations and actual measurements. Discussion: All three proposed seizure detection algorithms were shown to be suitable for application in implantable devices. The applied methodology for a platform-independent energy estimation was proven to be accurate by way of hardware implementation of the RNN algorithm. These findings show that seizure detection can be achieved using just a few channels with limited spatial distribution. The methodology proposed in this study can therefore be applied when designing new models for responsive neurostimulation.

Hardware Implementation of a Performance and Energy-optimized Convolutional Neural Network for Seizure Detection.

We present for the first time a µW-power convolutional neural network for seizure detection running on a low-power microcontroller. On a dataset of 22 patients a median sensitivity of 100% is achieved. With a false positive rate of 20.7 fp/h and a short detection delay of 3.4 s it is suitable for the application in an implantable closed-loop device.

A Comparison of Machine Learning Classifiers for Energy-Efficient Implementation of Seizure Detection.

The closed-loop application of electrical stimulation via chronically implanted electrodes is a novel approach to stop seizures in patients with focal-onset epilepsy. To this end, an energy efficient seizure detector that can be implemented in an implantable device is of crucial importance. In this study, we first evaluated the performance of two machine learning algorithms (Random Forest classifier and support vector machine (SVM)) by using selected time and frequency domain features with a limited need of computational resources. Performance of the algorithms was further compared to a detection strategy implemented in an existing closed loop neurostimulation device for the treatment of epilepsy. The results show a superior performance of the Random Forest classifier compared to the SVM classifier and the reference approach. Next, we implemented the feature extraction and classification process of the Random Forest classifier on a microcontroller to evaluate the energy efficiency of this seizure detector. In conclusion, the feature set in combination with Random Forest classifier is an energy efficient hardware implementation that shows an improvement of detection sensitivity and specificity compared to the presently available closed-loop intervention in epilepsy while preserving a low detection delay.

The Effect of the MEMS Measurement Platform Design on the Seebeck Coefficient Measurement of a Single Nanowire.

In order to study the thermoelectric properties of individual nanowires, a thermoelectric nanowire characterization platform (TNCP) has been previously developed and used in our chair. Here, we report on a redesigned platform aiming to optimize performance, mechanical stability and usability. We compare both platforms for electrical conductivity and the Seebeck coefficient for an individual Ag nanowire of the previously-used batch and for comparable measurement conditions. By this, the measurement performance of both designs can be investigated. As a result, whereas the electrical conductivity is comparable, the Seebeck coefficient shows a 50% deviation with respect to the previous studies. We discuss the possible effects of the platform design on the thermoelectric measurements. One reason for the deviation of the Seebeck coefficient is the design of the platform leading to temperature gradients along the bond pads. We further analyze the effect of bonding materials Au and Pt, as well as the effect of temperature distributions along the bond pads used for the thermovoltage acquisition. Another major reason for the variation of the measurement results is the non-homogeneous temperature distribution along the thermometer. We conclude that for the measurement of small Seebeck coefficients, an isothermal positioning of voltage-probing bond pads, as well as a constant temperature profile at the measurement zone are essential.

"Back on Track": A Mobile App Observational Study Using Apple's ResearchKit Framework.

BACKGROUND: In March 2015, Apple Inc announced ResearchKit, a novel open-source framework intended to help medical researchers to easily create apps for medical studies. With the announcement of this framework, Apple presented 5 apps built in a beta phase based on this framework. OBJECTIVE: The objective of this study was to better understand decision making in patients with acute anterior cruciate ligament (ACL) ruptures. Here, we describe the development of a ResearchKit app for this study. METHODS: A multilanguage observatory study was conducted. At first a suitable research topic, target groups, participating territories, and programming method were carefully identified. The ResearchKit framework was used to program the app. A secure server connection was realized via Secure Sockets Layer. A data storage and security concept separating personal information and study data was proposed. Furthermore, an efficient method to allow multilanguage support and distribute the app in many territories was presented. Ethical implications were considered and taken into account regarding privacy policies. RESULTS: An app study based on ResearchKit was developed without comprehensive iPhone Operating System (iOS) development experience. The Apple App Store is a major distribution channel causing significant download rates (>1.200/y) without active recruitment. Preliminary data analysis showed moderate dropout rates and a good quality of data. A total of 180 participants were currently enrolled with 107 actively participating and producing 424 completed surveys in 9 out of 24 months. CONCLUSIONS: ResearchKit is an easy-to-use framework and powerful tool to create medical studies. Advantages are the modular built, the extensive reach of iOS devices, and the convenient programming environment.

Design, Characterization and Sensitivity Analysis of a Piezoelectric Ceramic/Metal Composite Transducer.

This article presents experimental characterization and numerical simulation techniques used to create large amplitude and high frequency surface waves with the help of a metal/ceramic composite transducer array. Four piezoelectric bimorph transducers are cascaded and operated in a nonlinear regime, creating broad band resonant vibrations. The used metallic plate itself resembles a movable wall which can align perfectly with an airfoil surface. A phase-shifted operation of the actuators results in local displacements that generate a surface wave in the boundary layer for an active turbulence control application. The primary focus of this article is actuator design and a systematic parameter variation experiment which helped optimize its nonlinear dynamics. Finite Element Model (FEM) simulations were performed for different design variants, with a primary focus in particular on the minimization of bending stress seen directly on the piezo elements while achieving the highest possible deflection of the vibrating metallic plate. Large output force and a small yield stress (leading to a relatively small output stoke) are characteristics intrinsic to the stiff piezo-ceramics. Optimized piezo thickness and its spatial distribution on the bending surface resulted in an efficient stress management within the bimorph design. Thus, our proposed resonant transduction array achieved surface vibrations with a maximum peak-to-peak amplitude of 500 µ m in a frequency range around 1200 Hz.

Polydimethylsiloxane pressure sensors for force analysis in tension band wiring of the olecranon.

BACKGROUND: Several different surgical techniques are used in the treatment of olecranon fractures. Tension band wiring is one of the most preferred options by surgeons worldwide. The concept of this technique is to transform a tensile force into a compression force that adjoins two surfaces of a fractured bone. Currently, little is known about the resulting compression force within a fracture. OBJECTIVE: Sensor devices are needed that directly transduce the compression force into a measurement quality. This allows the comparison of different surgical techniques. Ideally the sensor devices ought to be placed in the gap between the fractured segments. METHODS: The design, development and characterization of miniaturized pressure sensors fabricated entirely from polydimethylsiloxane (PDMS) for a placement within a fracture is presented. The pressure sensors presented in this work are tested, calibrated and used in an experimental in vitro study. RESULTS: The pressure sensors are highly sensitive with an accuracy of approximately 3 kPa. A flexible fabrication process for various possible applications is described. The first in vitro study shows that using a single-twist or double-twist technique in tension band wiring of the olecranon has no significant effect on the resulting compression forces. CONCLUSIONS: The in vitro study shows the feasibility of the proposed measurement technique and the results of a first exemplary study.

Novel approach to dynamic knee laxity measurement using capacitive strain gauges.

PURPOSE: Knee joint laxities are observed in patients after severe trauma to the joint, resulting in ligament tears. Specifically, injuries to the anterior cruciate ligament may cause a significant instability. The degree of these laxities is essential in diagnostics and may affect which treatment option is suggested. METHODS: Polydimethylsiloxane (PDMS) strain gauges are proposed as a non-invasive, highly accurate and easy-to-use measurement method to quantify anterolateral and rotational laxities of the knee joint during active and passive motion. In this work, proof-of-concept measurements and a prototype of the proposed device are displayed. The measurements were taken using a knee test rig, which has specifically been designed for this purpose. This apparatus allows the simulation of isolated knee joint instabilities with a motor-controlled model of a human knee. RESULTS: The absolute sensitivity [Formula: see text] of an exemplary sensor was determined to be 2.038 [Formula: see text]; the relative sensitivity [Formula: see text] was 1.121 [Formula: see text]. Optimal positions of sensors to capture bone-to-bone displacement as projected displacement on the skin were identified. CONCLUSION: PDMS strain gauges are capable of measuring bone-to-bone displacements on the skin. We present an experimental in vitro study using an artificial knee test rig to simulate knee joint laxities and display the feasibility of our novel measurement approach.

Length Changes of the Anterolateral Ligament During Passive Knee Motion: A Human Cadaveric Study.

BACKGROUND: Persistent rotatory instability after anterior cruciate ligament (ACL) reconstruction may be a result of unaddressed insufficiency of the anterolateral structures. Recent publications about the anatomy of the anterolateral ligament (ALL) have led to a renewed interest in lateral extra-articular procedures, and several authors have proposed ALL reconstruction to supplement intra-articular ACL reconstruction. However, only limited knowledge about the biomechanical characteristics of the ALL exists. PURPOSE/HYPOTHESIS: The purpose of this study was to analyze length changes of the ALL during passive knee motion. The study hypothesis was that the ALL lengthens with knee flexion and internal tibial rotation. STUDY DESIGN: Controlled laboratory study. METHODS: The ALL of 6 cadaveric knees was dissected. Specimens were mounted in a specifically designed test rig that allowed unconstrained passive flexion/extension movement between 0° and 90° as well as external/internal tibial rotation of 25° at various flexion angles. Highly elastic, capacitive polydimethylsiloxane strain gauges were attached to the insertion sites of the ALL. Length changes were recorded continuously at flexion angles between 0° and 90° and during internal/external tibial rotation at 0°, 15°, 30°, 45°, 60°, 75°, and 90°. All measurements were calculated as the relative length change (%) of the ALL compared with 0° of flexion and neutral rotation. RESULTS: The mean relative length of the ALL significantly increased with increasing knee flexion (P < .001), with an estimated mean length change of +0.15% per degree. Both internal and external tibial rotation were independent determinants for length change; internal rotation significantly increased the length of the ALL (P < .001), whereas external rotation significantly decreased its length (P < .001). The mean length change with internal rotation increased with knee flexion, with a significantly greater length change at 90° compared with 0° (P = .048), 15° (P = .033), and 30° (P = .015). The maximum mean length change was +33.77% ± 9.62%, which was observed at 25° of internal rotation and 90° of flexion. CONCLUSION: The ALL is a nonisometric structure that tensions with knee flexion and internal tibial rotation. Length changes with internal rotation were greater at higher flexion angles, with the greatest length change of the ALL observed at 90° of flexion. CLINICAL RELEVANCE: The ALL can be considered a stabilizer against internal tibial rotation, especially at deep flexion angles. With regard to ALL reconstruction procedures, tensioning and fixation of the graft should be performed near 90° of flexion because graft tensioning near extension may cause excessive ligament strain with increasing knee flexion.

Dielectrophoretic investigation of Bi2Te3 nanowires-a microfabricated thermoelectric characterization platform for measuring the thermoelectric and structural properties of single nanowires.

In this article a microfabricated thermoelectric nanowire characterization platform to investigate the thermoelectric and structural properties of single nanowires is presented. By means of dielectrophoresis (DEP), a method to manipulate and orient nanowires in a controlled way to assemble them onto our measurement platform is introduced. The thermoelectric platform fabricated with optimally designed DEP electrodes results in a yield of nanowire assembly of approximately 90% under an applied peak-to-peak ac signal Vpp = 10 V and frequency f = 20 MHz within a series of 200 experiments. Ohmic contacts between the aligned single nanowire and the electrodes on the platform are established by electron beam-induced deposition. The Seebeck coefficient and electrical conductivity of electrochemically synthesized Bi2Te3 nanowires are measured to be -51 µV K(-1) and (943 ± 160)/(Ω(-1) cm(-1)), respectively. Chemical composition and crystallographic structure are obtained using transmission electron microscopy. The selected nanowire is observed to be single crystalline over its entire length and no grain boundaries are detected. At the surface of the nanowire, 66.1 ± 1.1 at.% Te and 34.9 ± 1.1 at.% Bi are observed. In contrast, chemical composition of 64.2 at.% Te and 35.8 at.% Bi is detected in the thick center of the nanowire.