Comparing classification techniques for identification of grasped objects.

BACKGROUND: This work presents a comparison and selection of different machine learning classification techniques applied in the identification of objects using data collected by an instrumented glove during a grasp process. The selected classifiers techniques can be applied to e-rehabilitation and e-training exercises for different pathologies, as in aphasic patients. METHODS: The adopted method uses the data from a commercial instrumented glove. An experiment was carried out, where three subjects using an instrumented glove had to grasp eight objects of common use. The collected data were submitted to nineteen different classification techniques (available on the scikit-learn library of Python) used in two classifier structures, with the objective of identifying the grasped object. The data were organized into two dataset scenarios: one with data from the three users and another with individual data. RESULTS: As a result of this work, three classification techniques presented similar accuracies for the classification of objects. Also, it was identified that when training the models with individual dataset the accuracy improves from 96 to 99%. CONCLUSIONS: Classification techniques were used in two classifier structures, one based on a single model and the other on a cascade model. For both classifier structure and scenarios, three of the classification techniques were selected due to the high reached accuracies. The highest results were obtained using the classifier structure that employed the cascade models and the scenario of individual dataset.

Hysteresis Compensation in a Tactile Device for Arterial Pulse Reproduction.

This paper describes a system for training healthcare practitioners in the identification of different arterial pulses. The driving system uses a linear solenoid in an open loop force control. Due to the large hysteresis it exhibited, a form of compensation was implemented, based on the classic Preisach model of hysteresis. Implementation of said model resulted in a significant reduction of force tracking error, demonstrating the feasibility of the chosen approach for the intended application.

An augmented reality interface to control a collaborative robot in rehab: A preliminary usability evaluation.

Human emotions can be seen as a valuable variable to explore in Human-Computer Interaction for effective, efficient, and satisfying interface development. The inclusion of appropriate emotional triggers in the design of interactive systems can play a decisive role in users' acceptance or rejection. It is well known that the major problem in motor rehabilitation is the high dropout rate resulting from the frustrated expectations given the typical slow recovery process and consequent lack of motivation to endure. This work proposes grouping a collaborative robot with one specific augmented reality equipment to create a rehabilitation system where some gamification levels might be added to provide a better and more motivating experience to patients. Such a system, as a whole, is customizable to adapt to each patient's needs on the rehabilitation exercises. By transforming a tedious exercise into a game, we expect to create an additional layer of enjoyment that can help in triggering positive emotions and stimulate users to continue the rehabilitation process. A pre-prototype was developed to validate this system's usability, and a cross-sectional study using a non-probabilistic sample of 31 individuals is presented and discussed. This study included the application of three standard questionnaires on usability and user experience. The analyses of these questionnaires show that the majority of the users found the system easy and enjoyable. The system was also analysed by a rehabilitation expert who gave a positive output regarding its usefulness, and positive impact on its use in the upper-limb rehabilitation processes. These results clearly encourage further development of the proposed system.

Accuracy of a digital skinfold system for measuring skinfold thickness and estimating body fat.

The use of skinfold thickness measurements to evaluate the distribution of subcutaneous adipose tissue and to predict body fat has recognised advantages. However, the different types of skinfold calliper available present limitations that make them unattractive and perhaps less used in daily practice. The purpose of the present study was to evaluate the accuracy and functionality of a new digital skinfold system, the Liposoft 2008+Adipsmeter V0 (LA), for measuring skinfold thickness and determining body fat proportion (%BF). Skinfold thickness measurements made by the LA were compared with those obtained with a Harpenden (H) calliper from two samples of adults (n 45) and older adults (n 56) in a university-based cross-sectional study. A comparison was also conducted between estimated %BF from skinfolds and dual-energy X-ray absorptiometry. Bland and Altman plots show that skinfolds measured by the LA and H calliper are in high agreement, with a mean difference of 0·3 (95% CI -3·1, 3·4) mm. In regard to the %BF estimated from LA and H skinfolds measurement, the LA produced a similar approximation to dual-energy X-ray absorptiometry %BF, with a mean difference of 0·2 (95% CI -0·8, 1·2) %, compared with %BF obtained with the H calliper. The LA system is an accurate instrumentation and represents an innovation in the evaluation of skinfold thickness and body composition based on anthropometric measurement.