Custom-made Sensors Research Articles

On-farm NIR sensor for milk analysis: Exploitation of bias monitoring and bias correction

Long-term studies have shown a bias drift over time in the prediction performance of near-infrared spectroscopy measurement systems. This bias drift generally requires extra laboratory reference measurements to detect and correct for this bias. Since these reference measurements are expensive and time consuming, there is a need for advanced methodologies for bias drift monitoring and correction without the need for taking extra samples. In this study, we propose and validate a method to monitor the bias drift and two methods to tackle it. The first method requires no extra measurements and uses a modified version of Partial Least Squares Regression to estimate and correct the bias. This method is based on the assumption that the mean concentration of the predicted component remains constant over time. The second method uses regular bulk milk measurements as a reference for bias correction. This method compares the measured concentrations of the bulk milk to the volume-weighted average concentrations of individual milk samples predicted by the sensor. Any difference between the actual and calculated bulk milk composition is then used to perform a bias correction on the predictions by the sensor system. The effectiveness of these methods to improve the component prediction was evaluated on data originating from a custom-built sensor that automatically measures the NIR reflectance and transmittance spectra of raw milk on the farm. We evaluate the practical use case where models for predicting the milk composition are trained upon installation of the sensor at the farm, and later used to predict the composition of subsequent samples over a period of more than 6 months. The effectiveness of the fully unsupervised method was confirmed when the mean concentration of the milk samples remained constant, while the effectiveness reduced when this was not the case. The bulk milk correction method was effective when all relevant samples for the component were measured by the sensor and included in the analyzed bulk milk, but is less effective when samples included in the bulk which are not measured by the sensor system. When the necessary conditions are met, these methods can be used to extend the lifetime of deployed prediction models by significantly reducing the bias on the predicted values.

A new approach to meteorological observations on remote polar glaciers using open-source internet of things technologies

Key regions of the world lack sufficient infrastructure to collect geophysical observations, often due to logistical challenges such as difficult accessibility and cost. With the advent of Internet-of-Things (IoT) technologies and low-cost electronics, it is possible today to build monitoring systems collecting spatially distributed, in-situ data with real-time connectivity to online servers for immediate and long-term usage at costs comparable to those of a single autonomous weather station. We present here a custom-built, modular system that collects quality data, and, that is, robust to adverse meteorological conditions and lack of energy. It integrates commercial and custom-built sensors connected to a node (main device) that manages power, data and radio communication. Data is sent to gateways and then to a server that parses, stores and quality controls the data. We deployed two networks in the vicinity of Ny-Ålesund in Svalbard, and operated from May 2021 to April 2022 to measure meteorological and glaciological variables. Our system collected reliable data and had sufficient power resources to survive 4–5 months of darkness during the polar night. Here, we present the design considerations and performance metrics, report our lessons learned from this challenging deployment, and suggest pathways for future improvements.

A Novel Computationally Efficient Approach to Identify Visually Interpretable Medical Conditions from 2D Skeletal Data

Timely identification and treatment of medical conditions could facilitate faster recovery and better health. Existing systems address this issue using custom-built sensors, which are invasive and difficult to generalize. A low-complexity scalable process is proposed to detect and identify medical conditions from 2D skeletal movements on video feed data. Minimal set of features relevant to distinguish medical conditions: AMF, PVF and GDF are derived from skeletal data on sampled frames across the entire action. The AMF (angular motion features) are derived to capture the angular motion of limbs during a specific action. The relative position of joints is represented by PVF (positional variation features). GDF (global displacement features) identifies the direction of overall skeletal movement. The discriminative capability of these features is illustrated by their variance across time for different actions. The classification of medical conditions is approached in two stages. In the first stage, a low-complexity binary LSTM classifier is trained to distinguish visual medical conditions from general human actions. As part of stage 2, a multi-class LSTM classifier is trained to identify the exact medical condition from a given set of visually interpretable medical conditions. The proposed features are extracted from the 2D skeletal data of NTU RGB + D and then used to train the binary and multi-class LSTM classifiers. The binary and multi-class classifiers observed average F1 scores of 77% and 73%, respectively, while the overall system produced an average F1 score of 69% and a weighted average F1 score of 80%. The multi-class classifier is found to utilize 10 to 100 times fewer parameters than existing 2D CNN-based models while producing similar levels of accuracy.

Guiding the design of superresolution tactile skins with taxel value isolines theory.

Tactile feedback is essential to make robots more agile and effective in unstructured environments. However, high-resolution tactile skins are not widely available; this is due to the large size of robust sensing units and because many units typically lead to fragility in wiring and to high costs. One route toward high-resolution and robust tactile skins involves the embedding of a few sensor units (taxels) into a flexible surface material and the use of signal processing to achieve sensing with superresolution accuracy. Here, we propose a theory for geometric superresolution to guide the development of tactile sensors of this kind and link it to machine learning techniques for signal processing. This theory is based on sensor isolines and allows us to compute the possible force sensitivity and accuracy in contact position and force magnitude as a spatial quantity before building a sensor. We evaluate the influence of different factors, such as elastic properties of the material, structure design, and transduction methods, using finite element simulations and by implementing real sensors. We empirically determine sensor isolines and validate the theory in two custom-built sensors with 1D and 2D measurement surfaces that use barometric units. Using machine learning methods to infer contact information, our sensors obtain an average superresolution factor of over 100 and 1200, respectively. Our theory can guide future tactile sensor designs and inform various design choices.

On-board cable attitude measurement and controller for outdoor aerial transportation

AbstractDeploying quadcopters for aerial transportation can be cost effective in impromptu material handling applications. However, such applications are limited mainly due to the requirement of onboard localization sensors and associated computation. The current work presents a human-controlled modality to successfully execute spontaneous outdoor flight of a quadcopter with a cable-suspended payload. Stable and smooth flights are achieved through an onboard integration of a custom-built sensor system and a controller to minimize payload oscillations. The feasibility of the proposed modality is demonstrated by conducting outdoor experiments and a case study in an unstructured environment.

Training with Agency-Inspired Feedback from an Instrumented Glove to Improve Functional Grasp Performance.

Sensory feedback from wearables can be effective to learn better movement through enhanced information and engagement. Facilitating greater user cognition during movement practice is critical to accelerate gains in motor function during rehabilitation following brain or spinal cord trauma. This preliminary study presents an approach using an instrumented glove to leverage sense of agency, or perception of control, to provide training feedback for functional grasp. Seventeen able-bodied subjects underwent training and testing with a custom-built sensor glove prototype from our laboratory. The glove utilizes onboard force and flex sensors to provide inputs to an artificial neural network that predicts achievement of “secure” grasp. Onboard visual and audio feedback was provided during training with progressively shorter time delay to induce greater agency by intentional binding, or perceived compression in time between an action (grasp) and sensory consequence (feedback). After training, subjects demonstrated a significant reduction (p < 0.05) in movement pathlength and completion time for a functional task involving grasp-move-place of a small object. Future work will include a model-based algorithm to compute secure grasp, virtual reality immersion, and testing with clinical populations.

Hafnium oxide layer-enhanced single-walled carbon nanotube field-effect transistor-based sensing platform

Single-walled carbon nanotube-based field effect transistors (SWCNT-FETs) are ideal candidates for fabricating sensors and have been widely used for chemical sensing applications. SWCNT-FETs have low selectivity because of the environmentally sensitive electronic properties of SWCNTs, and SWCNT-FETs also show a high noise signal and poor sensitivity because of charge trapping from Si–OH hydration of the SiO2/Si substrate on the SWCNTs. Herein, poly (4-vinylpyridine) (P4VP) was used for noncovalent attachment to SWCNTs and selective binding to copper ions (Cu2+). Importantly, the introduction of a hafnium-oxide (HfO2) layer through atomic layer deposition (ALD) overcame the charge trapping by SiO2 hydration and remarkably decreased the interference signal. The sensitivity of the P4VP/SWCNT/HfO2-FET sensor for Cu2+ was 7.9 μA μM−1, which was approximately 100 times higher than that of the P4VP/SWCNT/SiO2-FET sensor, and its limit of detection (LOD) was as low as 33 pmol L−1. Thus, the P4VP/SWCNT/HfO2-FET sensor is a promising candidate for the development of Cu2+-selective sensors and can be designed for the large-scale manufacturing of custom-made sensors in the future.

Thickness Measurement of Metallic Film Based on a High-Frequency Feature of Triple-Coil Electromagnetic Eddy Current Sensor

Previously, various techniques have been proposed for reducing the lift-off effect on the thickness measurement of the nonmagnetic films, including the peak-frequency feature and phase feature in the Dodd-Deed analytical formulation. To realize a real-time feedback response on the thickness monitoring, the phase term in the Dodd-Deeds formulation must be taken off the integration. Previous methods were based on the slow change rate of the phase term when compared to the rest of the term-the magnitude term. However, the change rate of the phase term is still considerable for a range of working frequencies. In this article, a high-frequency feature has been found. That is, the ratio between the imaginary and real parts of the phase term is proportional to the integral variable α under high frequencies. Based on this proportion relationship, the phase term has been taken out; and a thickness algorithm has been proposed. By combing the measured impedance from the custom-built sensor (three coils), the thickness of the metallic film can be reconstructed. Experiments have been carried out for the verification of the proposed scenario. Results show that the thickness of the metal film can be reconstructed with a small error of less than 2%, and immune to a reasonable range of lift-offs.

ACES

Many modern smart building applications are supported by wireless sensors to sense physical parameters, given the flexibility they offer and the reduced cost of deployment. However, most wireless sensors are powered by batteries today, and large deployments are inhibited by the requirement of periodic battery replacement. Energy harvesting sensors provide an attractive alternative, but they need to provide adequate quality of service to applications given the uncertainty of energy availability. We propose ACES, which uses reinforcement learning to maximize sensing quality of energy harvesting sensors for periodic and event-driven indoor sensing with available energy. Our custom-built sensor platform uses a supercapacitor to store energy and Bluetooth Low Energy to relay sensors data. Using simulations and real deployments, we use the data collected to continually adapt the sensing of each node to changing environmental patterns and transfer learning to reduce the training time in real deployments. In our 60-node deployment lasting 2 weeks, nodes stop operations only 0.1% of the time, and collection of data is comparable with current battery-powered nodes. We show that ACES reduces the node duty-cycle period by an average of 33% compared to three prior reinforcement learning techniques while continuously learning environmental changes over time.

RespiTrack: Patient-specific real-time respiratory tumor motion prediction using magnetic tracking

PurposeAn intraoperative real-time respiratory tumor motion prediction system with magnetic tracking technology is presented. Based on respiratory movements in different body regions, it provides patient and single/multiple tumor-specific prediction that facilitates the guiding of treatments.MethodsA custom-built phantom patient model replicates the respiratory cycles similar to a human body, while the custom-built sensor holder concept is applied on the patient’s surface to find optimum sensor number and their best possible placement locations to use in real-time surgical navigation and motion prediction of internal tumors. Automatic marker localization applied to patient’s 4D-CT data, feature selection and Gaussian process regression algorithms enable off-line prediction in the preoperative phase to increase the accuracy of real-time prediction.ResultsTwo evaluation methods with three different registration patterns (at fully/half inhaled and fully exhaled positions) were used quantitatively at all internal target positions in phantom: The statical method evaluates the accuracy by stopping simulated breathing and dynamic with continued breathing patterns. The overall root mean square error (RMS) for both methods was between 0.32pm 0.06~hbox {mm} and 3.71pm 0.79~hbox {mm}. The overall registration RMS error was 0.6pm 0.4~hbox {mm}. The best prediction errors were observed by registrations at half inhaled positions with minimum 0.27pm 0.02~hbox {mm}, maximum 2.90pm 0.72~hbox {mm}. The resulting accuracy satisfies most radiotherapy treatments or surgeries, e.g., for lung, liver, prostate and spine.ConclusionThe built system is proposed to predict respiratory motions of internal structures in the body while the patient is breathing freely during treatment. The custom-built sensor holders are compatible with magnetic tracking. Our presented approach reduces known technological and human limitations of commonly used methods for physicians and patients.

Localization of Signal Pattern Based UHF RFID Sensor Tags

This letter presents a prototype system that combines sensing and localization capabilities based on the radio frequency identification (RFID) technology in the ultrahigh-frequency (UHF) band. The system detects empty water tanks by recognizing a predefined transponder (tag) signal pattern at the reader. The shape of this pattern is determined by a custom-built sensor tag antenna that interacts in a unique way with an off-the-shelf microchip and the water filling level. The localization of this sensor tag is achieved by a 2-D direction of arrival (DoA) estimation, exploiting a $2\times2$ antenna array. The measurement results compare well with electromagnetic simulations. It is shown that the detection of the sensing state is sufficiently robust to cope with different tank positions, even in multipath environments. Furthermore, the system is robust in its localization performance, although the sensing state influences the wireless channel.

UHF RFID Sensor System Using Tag Signal Patterns: Prototype System

This letter presents a novel passive radio frequency identification (RFID) sensor prototype system in the ultrahigh frequency (UHF) range, which is based on the detection of fully controlled patterns of the backscattered transponder (tag) signals. The UHF RFID tag antenna itself is used as the sensing device. A custom-built sensor tag has been designed to monitor the water filling level of a canister (i.e., canister full or empty). The measurement results of the prototype system compare well with electromagnetic simulations and verify the novel sensor system concept. As predicted by the model, the prototype system solves the two main issues in antenna-based RFID sensor systems experienced so far, i.e., the impairment of system performance by modulating the sensor state information on the tag signal and the need for recurring system calibration due to changes in reader–tag distance.

A Modified Impedance-Frequency Converter for Inexpensive Inductive and Resistive Sensor Applications.

In this paper an exceptionally simple transducer is presented that is developed for experimental and custom-made sensors with inductive or resistive impedance output. It is based on a venerable 555 Integrated Circuit in a modified astable configuration. Due to single supply 5 V operation, it is directly compatible with most modern microcontroller systems, such as the popular Arduino platform. Various exemplary sensor characteristics are presented, including displacement, force, magnetic field, temperature and light sensing applications. While the transducer is not designed for high accuracy, it allows for fast and inexpensive application of various experimental sensors, such as magnetoelastic or GMI (Giant Magneto Impedance) sensors.

Barkhausen Noise Assessment of the Surface Conditions Due to Deep Hole Drilling and Their Influence on the Fatigue Behaviour of AISI 4140

In many technical fields, high-strength steels like AISI 4140 are commonly used for highly dynamically loaded parts. Increasing demands on weight, performance and efficiency of the automotive industry lead to increasing demands on material properties. For surface conditioning, optimised machining processes are capable of improving the fatigue performance without increasing the production cost significantly. This paper compares the influence of three different sulphur contents and three different feed rates on the fatigue behaviour of deep hole drilled AISI 4140 fatigue specimens. The specimens were characterised regarding their surface condition, hardness and microstructure, and afterwards they were tested under fatigue loading for performance assessment. These tests were accompanied with Barkhausen noise analysis. The Barkhausen noise signal was detected by a custom-built sensor that is capable of detecting the magnetic values on the bore wall. Using this technique, a load-independent estimation of fatigue damage was established.

Luminescence-Based Optical Sensors Fabricated by Means of the Layer-by-Layer Nano-Assembly Technique.

Luminescence-based sensing applications range from agriculture to biology, including medicine and environmental care, which indicates the importance of this technique as a detection tool. Luminescent optical sensors are required to be highly stable, sensitive, and selective, three crucial features that can be achieved by fabricating them by means of the layer-by-layer nano-assembly technique. This method permits us to tailor the sensors′ properties at the nanometer scale, avoiding luminophore aggregation and, hence, self-quenching, promoting the diffusion of the target analytes, and building a barrier against the undesired molecules. These characteristics give rise to the fabrication of custom-made sensors for each particular application.

Differential Intracochlear Sound Pressure Measurements in Human Temporal Bones with an Off-the-Shelf Sensor.

The standard method to determine the output level of acoustic and mechanical stimulation to the inner ear is measurement of vibration response of the stapes in human cadaveric temporal bones (TBs) by laser Doppler vibrometry. However, this method is reliable only if the intact ossicular chain is stimulated. For other stimulation modes an alternative method is needed. The differential intracochlear sound pressure between scala vestibuli (SV) and scala tympani (ST) is assumed to correlate with excitation. Using a custom-made pressure sensor it has been successfully measured and used to determine the output level of acoustic and mechanical stimulation. To make this method generally accessible, an off-the-shelf pressure sensor (Samba Preclin 420 LP, Samba Sensors) was tested here for intracochlear sound pressure measurements. During acoustic stimulation, intracochlear sound pressures were simultaneously measurable in SV and ST between 0.1 and 8 kHz with sufficient signal-to-noise ratios with this sensor. The pressure differences were comparable to results obtained with custom-made sensors. Our results demonstrated that the pressure sensor Samba Preclin 420 LP is usable for measurements of intracochlear sound pressures in SV and ST and for the determination of differential intracochlear sound pressures.

The HydroNet ASV, a Small-Sized Autonomous Catamaran for Real-Time Monitoring of Water Quality: From Design to Missions at Sea

In this paper, we describe the design, the development, and the sea trials of a novel small-sized autonomous surface vehicle (ASV) designed for monitoring the coastal water quality. The vehicle is characterized by the capability of measuring hydrocarbon and heavy metal concentrations directly onboard by means of custom-made miniaturized sensors. This capability, novel for an ASV, is combined with a winch-based sampling system specifically designed for small-sized vehicles. The sampling system can collect water samples up to 50 m in depth and measure the physical/water quality parameters of the water column. With these two features, the HydroNet ASV provides an autonomous, practical, real-time monitoring system, conceived to complement the current water monitoring practices in which samples are collected by a dedicated boat and analyzed in specialized laboratories at a later stage. The design process had the aim of realizing a vehicle capable of hosting the sampling system and the custom-made sensors that represent a unique payload in the world of small-sized ASVs. A twofold objective was pursued: realizing an ASV suited for monitoring missions in realistic scenarios (e.g., attention was paid to avoid water sample contamination), at the same time limiting the size for the ease of transportability and deployment. Severe constraints rose from these considerations and were addressed during the realization of the robot such as reduced length/weight (that limit the available space for the sensor payload) and low draft and protected propellers to allow the ASV to navigate in shallow waters with likely floating obstacles such as plastic bags. We report the design process aiming at a tradeoff between ease of transportability (small vehicle), available payload, and navigation performance in terms of achievable speed, endurance, and resistance to environmental disturbances (favored by larger ASV dimensions). We present sea trials of the realized vehicle validating the design choices. In particular, a long-range mission is discussed in which the robot executed a monitoring survey covering autonomously 12.5 km in front of Livorno, Italy, coast.

A novel augmented reality simulator for skills assessment in minimal invasive surgery.

Over the past decade, simulation-based training has come to the foreground as an efficient method for training and assessment of surgical skills in minimal invasive surgery. Box-trainers and virtual reality (VR) simulators have been introduced in the teaching curricula and have substituted to some extent the traditional model of training based on animals or cadavers. Augmented reality (AR) is a new technology that allows blending of VR elements and real objects within a real-world scene. In this paper, we present a novel AR simulator for assessment of basic laparoscopic skills. The components of the proposed system include: a box-trainer, a camera and a set of laparoscopic tools equipped with custom-made sensors that allow interaction with VR training elements. Three AR tasks were developed, focusing on basic skills such as perception of depth of field, hand-eye coordination and bimanual operation. The construct validity of the system was evaluated via a comparison between two experience groups: novices with no experience in laparoscopic surgery and experienced surgeons. The observed metrics included task execution time, tool pathlength and two task-specific errors. The study also included a feedback questionnaire requiring participants to evaluate the face-validity of the system. Between-group comparison demonstrated highly significant differences (<0.01) in all performance metrics and tasks denoting the simulator's construct validity. Qualitative analysis on the instruments' trajectories highlighted differences between novices and experts regarding smoothness and economy of motion. Subjects' ratings on the feedback questionnaire highlighted the face-validity of the training system. The results highlight the potential of the proposed simulator to discriminate groups with different expertise providing a proof of concept for the potential use of AR as a core technology for laparoscopic simulation training.

Lead me gently: Facilitating knowledge gain through attention-aware ambient learning displays

This empirical study reports an intervention to investigate identified research challenges on the evaluation and use of ambient displays in a learning context with the objective to gain insights into the interplay between display design, user attention, and knowledge acquisition. The main research questions were whether an attention-aware display design can capture the user's focus of attention and whether this has an influence on the knowledge gain. A display prototype corresponding to the main ambient display characteristics was designed, applied in a controlled authentic setting, and evaluated accordingly. The prototype presented information and guidelines for first responders in emergency situations, especially in cases of cardiac arrest. The prototype was enhanced with a custom-built sensor to measure user attention and trigger interruptive notifications. The study was conducted among 52 employees working at a university campus. Using an experimental research design, a treatment group exposed to an attention-aware display design was compared to a control group. The results provide evidence that such a display design can attract and retain attention in such a way that the acquisition of knowledge (i.e. the comprehension of the presented information) is effectively facilitated.

How Do Vision and Hearing Impact Pedestrian Time-to-Arrival Judgments?

To determine how accurate normally sighted male and female pedestrians were at making time-to-arrival (TTA) judgments of approaching vehicles when using just their hearing or both their hearing and vision. Ten male and 14 female subjects with confirmed normal vision and hearing estimated the TTA of approaching vehicles along an unsignalized street under two sensory conditions: (1) using both habitual vision and hearing and (2) using habitual hearing only. All subjects estimated how long the approaching vehicle would take to reach them (i.e., the TTA). The actual TTA of vehicles was also measured using custom-made sensors. The error in TTA judgments for each subject under each sensory condition was calculated as the difference between the actual and estimated TTA. A secondary timing experiment was also conducted to adjust each subject's TTA judgments for their "internal metronome." Error in TTA judgments changed significantly as a function of both the actual TTA (p < 0.0001) and sensory condition (p < 0.0001). Although no main effect for gender was found (p = 0.19), the way the TTA judgments varied within each sensory condition for each gender was different (p < 0.0001). Females tended to be as accurate under either condition (p ≥ 0.01), with the exception of TTA judgments made when the actual TTA was 2 seconds or less and 8 seconds or longer, during which the vision-and-hearing condition was more accurate (p ≤ 0.002). Males made more accurate TTA judgments under the hearing only condition for actual TTA values 5 seconds or less (p < 0.0001), after which there were no significant differences between the two conditions (p ≥ 0.01). Our data suggest that males and females use visual and auditory information differently when making TTA judgments. Although the sensory condition did not affect the females' accuracy in judgments, males initially tended to be more accurate when using their hearing only.