Egont Alexandre Schenkel

Identification of hand postures by force myography using an optical fiber specklegram sensor

The identification of hand postures based on force myography (FMG) measurements using a fiber specklegram sensor is reported. The microbending transducers were attached to the user forearm in order to detect the radial forces due to hand movements, and the normalized intensity inner products of output specklegrams were computed with reference to calibration positions. The correlation between measured specklegrams and postures was carried out by artificial neural networks, resulting in an overall accuracy of 91.3% on the retrieval of hand configuration.

Optical Fiber Specklegram Sensor for Measurement of Force Myography Signals

The development of an optical fiber specklegram sensor for the assessment of force myography signals is reported. The device consists of microbending transducers attached to the user forearm by means of Velcro straps. The muscular stimuli generated in response to hand movements cause the fibers to be mechanically pressed by the deformer structures, resulting in light modulation. The optical signals are acquired and processed according to specklegram analysis, by computing the normalized intensity inner product of speckles for two reference postures. Finally, the average values are correlated to the fingers configurations by means of artificial neural networks. The system was evaluated for a set of 11 static gestures, making it possible to detect subtle variations in the spatial distribution of forces impressed by the forearm flexor and extensor muscles. Moreover, the technique was tested for different subjects, yielding an average accuracy of 89.9% on the estimation of fingers configurations, even with the utilization of a reduced number of transducers.

Development of an optical fiber FMG sensor for the assessment of hand movements and forces

The development of an optical fiber specklegram sensor for detection of force myography (FMG) signals produced by hand forces and movements is presented. The optomechanical transducers are attached to the user forearm, causing variations on speckle field intensities, which are processed and then correlated to particular hand postures. The results indicated the viability to identify changes on hand configurations and forces, in which the force retrieval can be implemented via artificial neural networks after a previous calibration. The fiber FMG sensor can be further applied on human-system interfaces, as well as integrated to glove-based sensors in order to provide a robust detection of hand movements and forces.

Optical Classification of Quartz Lascas by Artificial Neural Networks

A gradation method based on quartz lascas (lumps) transparency level is proposed. The samples were irradiated by transmitting light, and the images histograms were processed by artificial neural networks. Additionally, the results were compared to conventional classification methods, including density and visual analysis. The network designed with backpropagation architecture using 4 hidden layers of 10 neurons yielded to a relative error <24% in relation to manual classification, indicating a good agreement to the miners criteria. Furthermore, the implementation of competitive learning with 5 neurons resulted in correct discrimination of samples regarding their optical characteristics with a completely non-subjective approach.

Design of a glove-based optical fiber sensor for applications in biomechatronics

The development of a glove-based sensor using flexible optical fiber microbending transducers for applications in biomechatronics is reported. The devices are attached onto the dorsal surface of monitored joint, resulting in the detection of variations of ~4° in angular displacements. In addition, the sensor system was evaluated for the real-time detection of hand movements, as well as for the assessment of grasping patterns during the manipulation of cylindrical objects. The preliminary experiments indicated that sensor could be suitable for practical applications in robot-assisted rehabilitation and diagnosis of impaired subjects.

Optical fiber tactile sensor based on fiber specklegram analysis

An optical fiber tactile sensor based on specklegram analysis is demonstrated. The system consists of microbending transducers attached to semi-rigid plates and distributed in form of array. The specklegrams produced by the multimode fibers are recorded and them the variations on fiber status, related to force magnitude, are assessed by specklegrams correlation. Finally, the spatial information was obtained by using a data fusion approach based on Bayesian inference. The sensor presented ∼0.08 N sensitivity and ∼1 mm resolution. Moreover, the data fusion technique provided the estimation of force distribution over a 30 × 30 mm2 area using only 3 fibers.

Initial development of a high-performance filtration device with Vapor-Phase Axial Deposition Process

The Vapor-Phase Axial Deposition Process was studied in order to evaluate its capacity for producing high porous nanomaterials, which could be used for filtration processes. An initial device was developed with this tecnique, and a micropore with diameter of 1,0 ± 0,1 μm was obtained, in a material combining nanosilica with nanotitania particles.

Study of activation energy and reaction parameters on the pozzolanic reaction using quartz and silica nanoparticles

In order to check the influence of temperature and morphology in pozzolanic reations, different types of silica were tested for a range of temperature. The pH measurements were utilized to plot the activate energy curve and check the advance of reaction. This range of data allows us to conclude that the better condition to conduct these tests is 75oC. Besides, fumed silica produced by Tecnosil presentes lower induction time and better fitting when compared to theoretical curve and when linearized by integer methods.

Use of Optical Fiber Sensor and Design of Experiments Statistical Methodology for Improving Dye Retention in a Quartz Ceramic Porous Material.

The fabrication process parameters of a quartz ceramic porous material are deriberately varied, and its dye retention after filtering an aqueous solution is evaluated with an optical fiber sensor. The parameters are analyzed in order of improving the retention, and determining the best fabrication conditions. According to the results, a reduction of 16.25% in relative dye concentration was obtained for the material sintered under 850oC for 120 minutes.

Study of the activation energy of silica nanoparticles

In order to check the influence of the environment during amorphous silica formation, different types of silica nanoparticles were tested according to their morphology, kinetics of reaction and reaction development. Samples of amorphous silica produced by Tecnocil and by VAD at LIQC/Unicamp were analysed by SEM, whereas their kinetics were studied by measuring the variance in pH. The VAD silica showed lower value for activation energy and higher homogeneity in terms of particle size and shape.