The sign language is widely used by deaf people around the globe. As the spoken languages, several sign languages do exist. The way sign language is learned by deaf people may have some details to be improved, but one can state that the existing learning mechanisms are effective when we talk about a deaf child, for example. The problem arises for the non-deaf persons that communicate with the deaf persons – the so-called listeners. If, for example, one couple has a new child that turns to be deaf, these two persons find a challenge to learn the sign language. In one hand, they cannot stop their working life, especially because of this sad news turns to be more costly, on the other hand, the existing mechanisms target the deaf-persons and are not prepared for the listeners. This paper proposes a new playful approach to help these listeners to learn the sign language. The proposal is a serious game composed of two modes: School-mode and Competition-mode. The first provides a school-like environment where the user learns the letter-signs and the second provides the user an environment used towards testing the learned skills. Behind the scenes, the proposal is based on two phases: 1 – the creation of a gestures library, relying on the Kinect depth camera; and 2 – the real-time recognition of gestures, by comparing what the depth camera information to the existing gestures previously stored in the library. A prototype system, supporting only the Portuguese sign language alphabet, was developed – the Kinect-Sign – and tested in a Portuguese Sign-Language school resulting in a joyful acceptance of the approach.

Serious games are usually applied to cognitive learning. However, the development of recent Natural User Interfaces (NUI) enables us to use serious games for a more holistic training. This project proposes to implement an educative game that teaches medical students how to carry out a neurological examination. The purpose is to improve not only the cognitive knowledge of the young physicians but also their sensorimotor skills. The game takes place in a virtual doctor office, which is modelled in 3D. Many possible clinical cases related to different neurologic diseases are simulated. The users are tested at two skill levels. First of all, they have to choose the right exam according to the patient’s symptoms. This initial stage is used to assess the theoretical expertise of the students. If the choice is correct, the physical examination begins. In order to evaluate the motor performance, a Kinect NUI is used. Such a device allows for the natural and intuitive control of the doctor’s avatar because there is a direct link between the individual’s and the avatar’s joints. Due to the fact that the real movements of the student are recorded with some accuracy, it is possible to gauge whether or not the medical gesture is properly applied to the virtual patient. In particular, the force and the precision of the movements are measured and compared to reference values. After each examination immediate feedback is given, so that the learners may rectify their choices and improve the quality of their motor skills at the next training session. Overall, the objective of this software is to enable the future physicians to practice a basic medical exam on a wide scope of neurologic disorders, without any risk for the patients and through an enjoyable approach.

The recent development of inorganic based composites as low-cost materials in reinforced concrete structural strengthening and precast thin-walled components, requires the creation of models that predict the mechanical behaviour of these materials. Textile Reinforced Mortar (TRM) shows complex stress–strain behaviour in tension derived from the heterogeneity of its constituent materials. This complexity is mainly caused by the formation of several cracks in the inorganic matrix. The multiple cracking leads to a decrease in structural stiffness. Due to the severe conditions of the serviceability limit state in structural elements, the prediction of the stress–strain curve is essential for design and calculation purposes. After checking other models, an empirical nonlinear approach, which is based on the crack control expression included in the Eurocode 2, is proposed in this paper. Following this scope, this paper presents an experimental campaign focused on 31 TRM specimens reinforced with four different reinforcing ratios. The results are analysed and satisfactorily contrasted with the presented non-linear approach.

The strengthening of reinforced concrete structures with laminates of fibre reinforced polymeric (FRP) matrix has received considerable attention, although there still is lack of information on the more adequate modelling of the interface between FRP composites and concrete. An experimental programme is described and was designed to: (i) characterise glass FRP-to-concrete interface by shear tests; (ii) analyse reinforced concrete T-beams with external GFRP plates. Double shear tests were carried out based on 15 cm cubes with GFRP bonded to two opposite faces. The concrete T-beams were 3.0 m long and 0.28 m high and were loaded till rupture in 4-point bending tests. The external reinforcement system showed great strength increment in relation to the non retrofitted T-beam, confirming to be an effective approach to the flexural strengthening of RC beams. The computational analysis was based on a three dimensional smeared crack model. In total, 22 computational analyses were made. Models with and without interface FE associated with Mohr–Coulomb failure criterion for the FRP-to-concrete interface were defined and different strength types of concrete were considered. The rigid interface does not predict the rupture of the T-beam with precision; however, the results obtained for low concrete strengths revealed that rigid interfaces can be assumed when conjugated with the fixed crack approach. Consequently, a slightly stiffer response of the beam is obtained. The maximum bond stresses obtained from Finite Element Analysis (FEA) revealed that the models with rigid interfaces developed lower bond stresses due to the lack of relative displacements between both materials. The effects of assuming either fixed or rotated crack approaches were also compared. The rotated crack conjugated to a fine mesh in the vicinity of the GFRP-to-concrete stress led to a very good estimation of the bond stresses along the interface. The prediction of the T-beam rupture was also estimated with better results when the rotated crack was used in the model. In general, the FEA predicted with very good results the de-bonding of the GFRP-to-concrete interface of T-beams externally bonded with GFRP composites.

The strengthening of reinforced concrete structures by means of externally bonded fibre reinforced polymers (FRPs) is now routinely considered and applied in the retrofit or strengthening of structures. FRP composites have received a considerable attention from civil engineers in recent years due to the high strength-weight and stiffness/weight ratios when compared to other materials. However, when FRP composites are bonded to a concrete surface, there is a persistent potential problem that the FRP plates may debond prematurely from the concrete. This is a very important issue for the engineers who have to focus on the computational modelling of this phenomenon. Some studies can be found in literature on computational modelling. However, there is very little information about the best modelling of the interface between FRP composites and concrete and this work is intended to help bridge this gap. The computational analysis presented here is based on three-dimensional software which assumes the smeared crack model, and the interface finite elements (FEs) used have a rupture criteria based on the Mohr-Coulomb criterion with tension cut-off. The definition of these FEs was based on double shear tests that were performed specifically for this purpose and they have shown that the debonding phenomenon can be predicted with some accuracy. In total, 10 double shear models were studied and the results were compared with the 21 experimental tests performed. The double shear tests consisted of applying loads to 2 layered GFRP laminates bonded to a 150 mm concrete cube with a bonded area of 150 × 80 mm (length × width). Double shear models with and without a gap interface were considered in order to emphasize the importance of modelling the GFRP-to-concrete interface with interface finite elements. The effect of the concrete strength on the interface performance was also considered. An externally bonded reinforcement (EBR) concrete T-beam strengthened with 2 GFRP layers is presented to illustrate the application of the method. The wet lay-up technique was used for the external reinforcement of a reinforced concrete T-beam and then tested under a four point bending test until rupture. The results are reported and differences between the numerical and the experimental results are discussed.

A crescente utilização de Unidades de Processamento Gráfico (GPUs) na computação de caráter geral levanta questões de desempenho e de escalabilidade. Para responder a estes requisitos de forma efetiva, cada vez mais se recorre à utilização colaborativa de vários GPUs num só sistema. Esta abordagem introduz, no entanto, novos desafios, tal como a decomposição do domínio do problema e a gestão da possível heterogeneidade dos dispositivos. Neste contexto assume particular relevância a proposta de abstrações que escondam a complexidade da programação destes sistemas. Existe já algum trabalho na área, mas este restringe-se ao paralelismo de dados. Por conseguinte, neste artigo abordamos a utilização de uma biblioteca de esqueletos algorítmicos, Marrow, para a exploração de paralelismo de tarefas em sistemas computacionais com estas características. Os resultados são promissores, apresentado a escalabilidade esperada nos sistemas testados.

Os sistemas de memória transacional distribuída atuais recorrem essencialmente à distribuição ou à replicação total para distribuir os seus dados pelos múltiplos nós do sistema. No entanto, estas estratégias de replicação de dados apresentam limitações. A distribuição não oferece tolerância a falhas e a replicação total limita a capacidade de armazenamento do sistema. Nesse contexto, a replicação parcial de dados surge como uma solução intermédia, que combina o melhor das duas anteriores com o intuito de mitigar as suas desvantagens. Esta estratégia tem sido explorada no contexto das bases de dados distribuídas, mas tem sido pouco abordada no contexto da memória transacional e, tanto quanto sabemos, nunca antes tinha sido incorporada num sistema de memória transacional distribuída para uma linguagem de propósito geral. Assim, neste artigo propomos e avaliamos uma infraestrutura para replicação parcial de dados para programas Java bytecode, que foi desenvolvida com base num sistema já existente de memória transacional distribuída. A modularidade da infraestrutura que apresentamos permite a implementação de múltiplos algoritmos e, por conseguinte, avaliar em que contextos de utilização (workloads, número de nós, etc.) a replicação parcial se apresenta como uma alternativa viável a outras estratégias de replicação de dados.

O controlo da concorrência no acesso a estado partilhado assume actualmente um papel de destaque no desenvolvimento de software. Trabalhos recentes propõem que tal gestão seja expressa ao nível dos dados, em alternativa à usual centralidade no código. A principal vantagem é o acoplamento da gestão da concorrência com a declaração dos dados, eliminando desse modo a descentralização dos erros de concorrência, facilitando a sua correção. No entanto, as abordagens centradas nos dados existentes pecam por não garantirem a ausência de deadlocks em todos os cenários e/ou exigirem do programador a agregação explícita dos recursos que devem ser avaliados atomicamente. A nossa proposta colmata ambas estas limitações. O programador anota isoladamente que zonas de memória requerem acesso exclusivo, sendo que uma análise estática infere quais dessas devem ser agrupadas e adquiridas atomicamente, e garante que o código gerado é ausente de deadlocks. De modo a aferir-se a eficiência da nossa solução, comparamos o seu desempenho e a sua produtividade relativamente à memória transacional e outras abordagens centrada nos dados.

A decomposição correta de um problema paralelo com base na hierarquia de memória onde irá executar pode levar a ganhos de desempenho significativos durante execução do mesmo. No entanto, os subsistemas de memória das arquiteturas multicore modernas apresentam variadas configurações, em termos das suas organizações hierárquicas e da capacidade dos seus diversos níveis de memória. Existem diversas abordagens que permitem adequar a execução de uma aplicação à estratificação hierárquica da memória,. Contudo estas exigem do programador um conhecimento profundo da arquitetura alvo e de programação paralela em geral. A abordagem apresentada neste artigo contrasta com as demais, transpondo esta responsabilidade para o sistema de execução, colocando sobre a sua alçada a decomposição hierárquica da computação. Nessa medida, ao programador cabe apenas expressar de forma genérica os algoritmos de subdivisão do domínio do problema. Avaliamos o desempenho da nossa abordagem relativamente a outra baseada na usual decomposição horizontal do domínio do problema. Os resultados são bons, apresentando ganhos de performance em aplicações que usufruem do tipo de otimização efetuada e desempenhos equiparáveis nas restantes.