Search

Between two-dimensional (2D) semiconductors, the family of transition metal dichalcogenides (TMDs) exhibits promising optical properties, such as tightly bound excitons and single photon emitting centers, with favourable integration in current nano-photonic architectures. High refractive-index dielectrics have recently gained attention as they exhibit optical Mie resonances, offering a novel platform to engineer light-matter interaction down to sub-wavelength dimensions. In our work, we integrated single and double layer TMDs WSe 2 with Mie-resonant gallium phosphide (GaP) dimer nano-antennas and demonstrate the enhancement of the light-matter interaction of 2D excitons and relative single photon emitters.

In this paper, a millimeter-wave radar operating at the frequency of 120 GHz is employed as an aid for people with neurodegenerative disorders. In detail, an interferometric approach is exploited for detecting the movement of the head and the eyes blinking. As a matter of fact, these actions are among the very few ways to enable the communication of a person with the surrounding environment in case of heavy neurodegenerative disorder. The theoretical principles and the relevant aspects of the proposed solution are shown by simulating the entire scenario, whereas the effectiveness of the system is experimentally demonstrated.

The BepiColombo mission (ESA/JAXA) is devoted to the exploration of Mercury, the innermost planet of the Solar system. Among the experiments onboard of the spacecraft, the Mercury Orbit Radio Science Experiment (MORE) aims to gather information about the surface and the internal structure of the planet, and to carry out fundamental physics tests, thanks to a state-of- the-art tracking system. During the cruise phase of BepiColombo, a general relativity test is performed in order to determine the Eddington parameter. Due to the proximity to the Sun, the solar irradiance random fluctuations have a dynamical effect that cannot be neglected. In this work a complete numerical simulation of the MORE fundamental physics experiment in cruise is carried out, taking into account the irradiance fluctuations. A precise orbit determination simulation has been used in order to assess the attainable accuracy on the determination of the Eddington parameter.

This paper describes the calibration experiments that took place at Loch Goil, Scotland, as part of the HR-LFSAS project. The aim of the trials was to calibrate a 2D low frequency array and to measure its beampattern. One of the main challenges of the calibration was the sampling frequency at the transmitting side introducing strong spectrum periodisation in the recorded signal. The artefacts impact heavily the received signal at frequencies close to the Nyquist frequency. Utilising the high sampling frequency of the receive hydrophone, we propose here an algorithm to reconstruct the pulse based on a down-sampling and up-sampling procedure. We also derive a metric based on spectrum relative energy to assess the quantitative efficiency of a DAC (Digital-to-Analog Convertor) system as a function of the ratio between its sampling frequency and the pulse maximum frequency. Finally, we show that under certain conditions, a system can be calibrated beyond the Nyquist frequency imposed by the DAC.

This paper presents the modular design of a transmitter array for a High Resolution Low Frequency Synthetic Aperture Sonar (HRLFSAS). The proposed solution is equipped with a 2D transmitting array with independently driven elements, capable of delivering wideband signals. The transmitter is coupled with a receiver array to perform high-resolution SAS (Synthetic Aperture Sonar) acoustic imaging. This sonar has been designed to be modular and scalable with high source level, reduced reverberation and completely programmable aperture and steering capabilities. The design and realization of this system has required a considerable effort from CMRE, as the solution presented is unique in the sonar panorama: the integration of nearly 200 independent transducers in a considerably compact system (being a low frequency system) will allow carrying on advanced studies on object and bottom penetrating sonars. This paper provides the description of the system transmit 2D array characteristics in terms of its modular design.

Robot-assisted rehabilitation in children and young adults with Cerebral Palsy (CP) is expected to lead to neuroplasticity and reduce the burden of motor impairments. For a lower-limb exoskeleton to perform well in this context, it is essential that the robot be "transparent" to the user and produce torques only when voluntarily-generated motor outputs deviate significantly from the target trajectory. However, the development of transparent operation modes and assistance-as-need control schema are still open problems with several implementation challenges. This paper presents a theoretical approach and provides a discussion of the key issues pertinent to designing a transparent operation mode for a lower-limb exoskeleton suitable for children and young adults with CP. Based on the dynamics of exoskeletons as well as friction models and human-robot interaction models, we propose a control strategy aimed to minimize human-machine interaction forces when subjects generate motor outputs that match the target trajectory. The material is presented as a conceptual framework that can be generalized to other exoskeleton systems for overground walking.

The wide diffusion of Distributed Generation, the need to improve the quality of service and the progressive evolution of technology have led to an evolution of the distribution network carrying out the e-distribuzione’s idea of Smart Grids. Moving in this direction, e-distribuzione built up two laboratory located in the cities of Milan and Bari to carry out research and development activity on smart grids. In particular, real time digital simulators and emulators are used to develop and test new equipment, instruments and sensors for digitizing the grid and ensuring its integration into the e-distribuzione ecosystem.This paper illustrates the Smart Grid Labs of e-distribuzione, where new infrastructures for remote control, automation and advanced management of the Medium Voltage (MV) and Low Voltage (LV) networks are designed and implemented. Finally, the paper shows the innovative projects in which the new Smart Grid devices are involved.

This paper presents a minimum complexity hand movement recognition algorithm based on Average Threshold Crossing (ATC) technique. It exploits the number of threshold-crossing events, generated by a full-custom acquisition board, from the surface ElectroMyoGraphic (sEMG) signals of three forearm muscles to detect four different movements of the wrist: flexion, extension, abduction and grasp. A Support Vector Machine (SVM) model has been trained with the signals acquired from ten subjects, who repeated ten times each gesture. To avoid correlation between training and testing dataset, the Leave One Subject Out (LOSO) cross-validation technique has been chosen. The average ATC classifier's accuracy is 92.87 %, only 5.34 % below the results obtained feeding the same model with the sEMG features extracted from the raw sampled signals. The total latency of the algorithm, from the acquisition to the prediction, is 160 ms. Power consumption was considered too: with less than the power budget for one sampled sEMG channel, it is possible to acquire and transmit (through a Bluetooth low energy module) the event-driven data of four sEMG channels, with an effective data rate of only 28 B/s. Obtained performance makes this technique suited for wearable systems or Internet-of-Things (IoT) applications.

Online data provide a way to monitor how users behave in social systems like social networks and online games, and understand which features turn an ordinary individual into a successful one. Here, we propose to study individual performance and success in Multiplayer Online Battle Arena (MOBA) games. Our purpose is to identify those behaviors and playing styles that are characteristic of players with high skill level and that distinguish them from other players. To this aim, we study Defense of the ancient 2 (Dota 2), a popular MOBA game. Our findings highlight three main aspects to be successful in the game: (i) players need to have a warm-up period to enhance their performance in the game; (ii) having a long in-game experience does not necessarily translate in achieving better skills; but rather, (iii) players that reach high skill levels differentiate from others because of their aggressive playing strategy, which implies to kill opponents more often than cooperating with teammates, and trying to give an early end to the match.

An acquisition system exploiting event-based approach applied to surface ElectroMyoGraphy (sEMG) was developed, with the goal of reducing the device size and minimizing the power consumption. The device has a modular four-channels architecture; each single acquisition module is characterized by a power consumption of 0.635 mW, a body area of 414 mm2 and a weight of 2 g, making it a good fit for wearable applications. In-vivo experiments have been conducted on 11 subjects, during isometric contraction with increasing load of the Biceps Brachii and Vastus Lateralis muscles. The device was able to discriminate different muscular activation levels, through the average number (over time) of threshold-crossing events, providing promising results in terms of correlation with the weight (median values of 0.97 and 0.95 respectively), compared to standard sEMG parameters described in literature.

We present a deep-learning framework for real-time multiple spatio-temporal (S/T) action localisation, classification and early prediction. Current state-of-the-art approaches work offline and are too slow to be useful in real- world settings. To overcome their limitations we introduce two major developments. Firstly, we adopt real-time SSD (Single Shot MultiBox Detector) convolutional neural networks to regress and classify detection boxes in each video frame potentially containing an action of interest. Secondly, we design an original and efficient online algorithm to incrementally construct and label `action tubes' from the SSD frame level detections. As a result, our system is not only capable of performing S/T detection in real time, but can also perform early action prediction in an online fashion. We achieve new state-of-the-art results in both S/T action localisation and early action prediction on the challenging UCF101-24 and J-HMDB-21 benchmarks, even when compared to the top offline competitors. To the best of our knowledge, ours is the first real-time (up to 40fps) system able to perform online S/T action localisation and early action prediction on the untrimmed videos of UCF101-24., 10 pages 3 figures, ICCV 2017, Added link to new annotations of ucf101-24

Platforms with different computation resource, e.g. CPUs and FPGAs, become one of the first choices to deploy performance-requiring embedded applications. On this technology, functionalities can be implemented either as hardware (HW) or software (SW) components. Here, we extend the MultiPar methodology to support the selection of optimal partitioning solutions with respect to system properties. We show the feasibility of the proposed methodology and validate the composition rules for properties used in the partitioning decision process.

Current object detection approaches predict bounding boxes, but these provide little instance-specific information beyond location, scale and aspect ratio. In this work, we propose to directly regress to objects' shapes in addition to their bounding boxes and categories. It is crucial to find an appropriate shape representation that is compact and decodable, and in which objects can be compared for higher-order concepts such as view similarity, pose variation and occlusion. To achieve this, we use a denoising convolutional auto-encoder to establish an embedding space, and place the decoder after a fast end-to-end network trained to regress directly to the encoded shape vectors. This yields what to the best of our knowledge is the first real-time shape prediction network, running at ~35 FPS on a high-end desktop. With higher-order shape reasoning well-integrated into the network pipeline, the network shows the useful practical quality of generalising to unseen categories similar to the ones in the training set, something that most existing approaches fail to handle., Comment: 16 pages including appendix; Published at CVPR 2017

The study herein summarized was focused on the development of a method to derive reliable estimates of the quality of movement of stroke survivors via the analysis of wearable sensor data. Data was collected from 34 subjects while they performed a battery of functional movements that are part of a standard clinical assessment. The quality of movement was assessed using the Functional Ability Scale, a validated clinical scale based on visual observation of movement patterns by a clinical expert. Two wearable sensors were positioned on the stroke-affected wrist and the sternum, respectively. Wearable sensor data was processed to derive data features that were in turn used as input to a regression-based Random Forest algorithm that was trained using a leave-one-subject-out method. The Random Forest algorithm generated estimates of the Functional Ability Scale scores. We found out that the estimates generated via analysis of the wearable sensor data were highly correlated (R2=0.97) with the scores generated by the clinical expert.

Photonic integration is as an enabling technology for photonic quantum science, providing great experimental scalability, stability, and functionality. Although the increasing complexity of quantum photonic circuits has allowed proof-of-principle demonstrations of quantum computation, simulation, and metrology[1], further development is severely limited by the on-chip photon flux that can be made available from external quantum light sources[2]. Overcoming such limitations would allow a significant scaling of quantum photonic experiments, and enable quantum-level investigation of many physical processes observable on-chip through nanophotonic and nanoplasmonic structures (e.g., Kerr, optomechanical, single-photon nonlinearities). Towards such goals, we have developed a scalable, heterogeneous III-V/Si 3 N 4 integration platform for quantum photonic circuits based on passive Si 3 N 4 waveguides which directly incorporate nanophotonic single-photon sources based on self-assembled InAs quantum dots (QDs)[3]. InAs quantum dots constitute the most promising solid-state triggered single-photon sources to date[4], while SI3N4 waveguides offer low-loss propagation, tailorable dispersion and high Kerr nonlinearities which can be used for linear and nonlinear optical signal processing down to the quantum level. In our platform, the building blocks of which are shown in Fig. 1(a), active GaAs waveguide-based geometries containing InAs QDs are designed to efficiently capture QD-emitted single-photons. Captured photons, confined within the GaAs core, are then transferred with high efficiency into a passive Si 3 N 4 waveguide network via adiabatic mode transformers. Figure 1(b) shows an example device fabricated with our platform: a GaAs microring resonator containing InAs quantum dots, evanescently coupled to a GaAs bus waveguide, which is in turn coupled to an underlying Si 3 N 4 waveguide through adiabatic mode-transformers. The photoluminescence spectrum for this device, in Fig. 1(b), shows that a single QD exciton near 1125 nm, coupled to a microring whispering-gallery mode, acts as a source of single-photons that are launched directly into the Si3N4 waveguide. This geometry also allows us to effectively control the QD spontaneous emission decay lifetime by spectrally detuning the WGM with respect to the QD, as shown in Fig. 1(d).

Coherent control of optical waves by scattering from 2D plasmonic surfaces is a very active field with the goal of wavefront shaping an incoming light beam via nanostructured surfaces. Here we report gold hyperuniform surfaces fabricated by electron beam lithography and designed to interact with visible radiation. We observe omnidirectional light diffraction in the far-field indicating successful k-space design with rotational symmetry, and spontaneous emission directional control of near-field coupled emitters.