This work aims at presenting a novel Computer Vision approach in the development of a real-time, web-camera based, British Sign Language recognition system. A literature review focused on current (1) state of sign language recognition systems and (2) techniques used is conducted. This review process is used as a foundation on which a Convolutional Neural Network (CNN) based system is designed and then implemented. A bespoke British Sign Language dataset -containing 11,875 images - is then performed to train and test the CNN which is used for the classification of human hand performed gestures. Finally, the CNN architecture recognized 19 static British Sign Language gestures, incorporating both single and double-handed gestures. During testing, the system achieved an average recognition accuracy of 89%.

This paper introduces an optical based three axis force sensor which can be integrated with the robot arm of the EU project STIFF-FLOP (STIFFness controllable Flexible and Learnable Manipulator for Surgical Operations) in order to measure applied external forces. The structure of the STIFF-FLOP arm is free of metal components and electric circuits and, hence, is inherently safe near patients during surgical operations. In addition, this feature makes the performance of this sensing system immune against strong magnetic fields inside magnetic resonance (MR) imaging scanners. The hollow structure of the sensor allows the implementation of distributed actuation and sensing along the body of the manipulator. In this paper, we describe the design and calibration procedure of the proposed three axis optics-based force sensor. The experimental results confirm the effectiveness of our optical sensing approach and its applicability to determine the force and momentum components during the physical interaction of the robot arm with its environment.

In the wind turbine manufacturing industry, the bolt-nut joint tightening process is one of the core processes in the full production chain and concerned with assembling the hub body, the pitch system and the bearing unit. This operation is currently executed manually with the aid of different tools and gauges; the main disadvantages are a relatively high degree of variability and the necessity to repeat this task several times during a production run to achieve a satisfactory, final tightening torque within a specified angle range. Moreover, the bolt tightening process includes various uncertainties such as the presence of friction forces and the use of different bolt sizes with different stiffness values which make it highly nonlinear and uncertain resulting in a challenging control problem. To facilitate the development of an effective control strategy, we study the bolt tightening process and propose 4 tightening stages, namely, bolt-nut alignment, partial and full engagement and final bolt tightening. Based on the characteristics of each stage, a fuzzy controller is designed for each stage to realize the respective control objectives. A fuzzy error detector incorporating the knowledge of each stage is proposed for early error detection, making use of the input from a torque and encoder (angular position) sensor. Errors can be detected in each stage to interrupt the process and prevent any damage to the system.

In the last few years, much effort has been devoted to the development of wearable sensing systems able to monitor physiological, behavioral, and environmental parameters. Less has been done on the accurate testing and assessment of this instrumentation, especially when considering devices thought to be used in harsh environments by subjects or operators performing intense physical activities. This paper presents methodology and results of the evaluation of wearable physiological sensors under these conditions. The methodology has been applied to a specific textile-based prototype, aimed at the real-time monitoring of rescuers in emergency contexts, which has been developed within a European funded project called ProeTEX. Wearable sensor measurements have been compared with the ones of suitable gold standards through Bland-Altman statistical analysis; tests were realized in controlled environments simulating typical intervention conditions, with temperatures ranging from 20 °C to 45 °C and subjects performing mild to very intense activities. This evaluation methodology demonstrated to be effective for the definition of the limits of use of wearable sensors. Furthermore, the ProeTEX prototype demonstrated to be reliable, since it produced negligible errors when used for up to 1 h in normal environmental temperature (20 °C and 35 °C) and up to 30 min in harsher environment (45 °C).

Assessment of human activity and posture with triaxial accelerometers provides insightful information about the functional ability: classification of human activities in rehabilitation and elderly surveillance contexts has been already proposed in the literature. In the meanwhile, recent technological advances allow developing miniaturized wearable devices, integrated within garments, which may extend this assessment to novel tasks, such as real-time remote surveillance of workers and emergency operators intervening in harsh environments. We present an algorithm for human posture and activity-level detection, based on the real-time processing of the signals produced by one wearable triaxial accelerometer. The algorithm is independent of the sensor orientation with respect to the body. Furthermore, it associates to its outputs a "reliability" value, representing the classification quality, in order to launch reliable alarms only when effective dangerous conditions are detected. The system was tested on a customized device to estimate the computational resources needed for real-time functioning. Results exhibit an overall 96.2% accuracy when classifying both static and dynamic activities.

Financed by the European Commission, a consortium of 23 European partners, consisting of universities, research institutions, industries, and organizations operating in the field of emergency management, is developing a new generation of "smart" garments for emergency-disaster personnel. Garments integrate newly developed wearable and textile solutions, such as commercial portable sensors and devices, in order to continuously monitor risks endangering rescuers' lives. The system enables detection of health-state parameters of the users (heart rate, breathing rate, body temperature, blood oxygen saturation, position, activity, and posture) and environmental variables (external temperature, presence of toxic gases, and heat flux passing through the garments), to process data and remotely transmit useful information to the operation manager. The European-integrated project, called ProeTEX (Protection e-Textiles: Micro-Nano-Structured fiber systems for Emergency-Disaster Wear) started on February, 2006 and will end on July, 2010. During this 4.5 years period, three subsequent generations of sensorized garments are being released. This paper proposes an overview of the project and gives a description of the second-generation prototypes, delivered at the end of 2008.

The current state of the art in wearable electronics is the integration of very small devices into textile fabrics, the so-called ¿smart garment.¿ The ProeTEX project is one of many initiatives dedicated to the development of smart garments specifically designed for people who risk their lives in the line of duty such as fire fighters and Civil Protection rescuers. These garments have integrated multipurpose sensors that monitor their activities while in action. To this aim, we have developed an algorithm that combines both features extracted from the signal of a triaxial accelerometer and one ECG lead. Microprocessors integrated in the garments detect the signal magnitude area of inertial acceleration, step frequency, trunk inclination, heart rate (HR), and HR trend in real time. Given these inputs, a classifier assigns these signals to nine classes differentiating between certain physical activities (walking, running, moving on site), intensities (intense, mild, or at rest) and postures (lying down, standing up). Specific classes will be identified as dangerous to the rescuer during operation, such as, ¿subject motionless lying down¿ or ¿subject resting with abnormal HR.¿ Laboratory tests were carried out on seven healthy adult subjects with the collection of over 4.5 h of data. The results were very positive, achieving an overall classification accuracy of 88.8%.

This paper describes the dynamic control of a 3 degree of freedom (DOF) ringer emulating a human finger for reaching a desired fingertip position in space. The control consists of a neural network (NN) which provides the necessary three torques to the phalanges granting a smooth "natural" speed profile of the fingertip motion. To obtain the results, we face these problems: 1) the elimination of the redundancy due to the third joint; 2) the mathematical description of a natural movement; 3) the calculation of the torques for executing movement; and 4) the optimization of the NN's structure. Assuming a "cognitive" approach well-established in the literature, the first and the second points are solved by adopting an extension of the minimum jerk theory. The classic Lagrange equations are applied to compute the three-motor torque. Finally, a multilayer perceptron (MLP) NN is trained to move the device in a natural manner. The generalization capabilities of the NN are checked on new never-seen movements, and different MLP architectures are compared on the basis of indexes representing the motor performance. The results suggest we should pursue this approach for multifinger hand in order to achieve a natural NN prosthetic/robotic dynamic control system.