1,721,036 research outputs found
NONLINEAR BLIND SOURCE SEPARATION BY SPLINE NEURAL NETWORKS
No full textn this paper a new neural network model for blind demixing of nonlinear mixtures is proposed. We address the use of the adaptive spline neural network recently introduced for supervised and unsupervised neural networks. These networks are built using neurons with flexible B-spline activation functions and in order to separate signals from mixtures, a gradient-ascending algorithm which maximizes the outputs entropy is derived. In particular a suitable architecture composed by two layers of flexible nonlinear functions for the separation of nonlinear mixtures is proposed. Some experimental results that demonstrate the effectiveness of the proposed neural architecture are presented
An optimization procedure based on kinematics analysis for the design parameters of a 4-UPU parallel manipulator
No full textThe paper presents the kinematics analysis of a 4-UPU fully parallel manipulator and a numerical procedure for its optimization. The 4-UPU kinematics is featured by four UPU legs, moreover the prismatic joint of each leg is supposed to be actuated. The end-effector has 4 degrees of freedom (Dofs), the three translations and the rotation along the direction perpendicular to the base platform. Singularity configurations have been analytically determined and the analysis of both actuation and constraint Jacobian by screw theory has been performed. Finally, a numerical procedure for the optimization of the design parameters of the manipulator is presented. The optimization procedure has been carried out using adimensional parameters in order to generalize the obtained results. The presented optimization procedure allows to define the ratios between the geometrical features of the manipulator which maximize a generic performance parameter in a designed workspace by avoiding all the possible singularity conditions. The obtained results are useful for the design of all the manipulators based on the 4-UPU kinematics. In particular, the case study of the design of a wearable fingertip haptic device is presented and discussed
Artificial Neural Network with Adaptive Multidimensional Spline Activation Functions
Full text linkThis work concerns a new kind of neural structure that involves a multidimensional adaptive activation function. The proposed architecture, based on multidimensional cubic spline, allows to collect information from the previous network layer in aggregate form. In other words the number of network connections (structural complexity) can be very low respect to the problem complexity. This fact, as experimentally demonstrated in the paper, improve the network generalization capabilities and speed up the convergence of the learning process. A specific learning algorithm is derived and experimental results demonstrate the effectiveness of the proposed architecture
Neural equalizer with adaptive multidimensional spline activation functions
No full textThis paper presents a new neural architecture suitable for digital signal processing application. The architecture, based on adaptable multidimensional activation functions, allows one to collect information from the previous network layer in aggregate form. In other words the number of network connections (structural complexity) can be very low respect to the problem complexity. This fact, as experimentally demonstrated in the paper, improve the network generalization capabilities and speed up the convergence of the learning process. A specific learning algorithm is derived and experimental results, on channel equalization, demonstrate the effectiveness of the proposed architecture
Design Requirements of Generic Hand Exoskeletons and Survey of Hand Exoskeletons for Rehabilitation, Assistive, or Haptic Use
No full textMost current hand exoskeletons have been designed specifically for rehabilitation, assistive, or haptic applications to simplify the design requirements. Clinical studies on poststroke rehabilitation have shown that adapting assistive or haptic applications into physical therapy sessions significantly improves the motor learning and treatment process. The recent technology can lead to the creation of generic hand exoskeletons that are application-agnostic. In this paper, our motivation is to create guidelines and best practices for generic exoskeletons by reviewing the literature of current devices. First, we describe each application and briefly explain their design requirements, and then list the design selections to achieve these requirements. Then, we detail each selection by investigating the existing exoskeletons based on their design choices, and by highlighting their impact on application types. With the motivation of creating efficient generic exoskeletons in the future, we finally summarize the best practices in the literature
Experimental identification of faults in a railway pantograph mechanism in presence of nonlinear damping (Aimeta 2017)
No full textPredictability may transform maintenance from a cost centre into a profit unit. Industries are indeed investing in innovative technologies to implement the purpose. In this paper we introduce a project meant to develop a robotic tool for the assessment of railway pantographs through a vibration based detection of faults. Nonlinearities such as intense dependence of damping on speed and non-negligible friction in joints make the applicaton particularly challenging. These phenomena could in fact drastically modify the behavior with respect to simpler linear models predictions. Experimental frequency response functions (FRFs) were hence derived for several input amplitudes so as to estimate at which extent modal parameters are related to the energy level. To discriminate the sources of nonlinearity FRFs were carried out even after the damper had been removed. Phase plots shape alteration was then observed for a couple of frequencies and as many amplitudes to delve deeper into dynamics changes. A first indication on the suitability of FRFs to detect faults is pursued considering the absence of the dashpot as an example of defect, i.e. a severe leakage, comparing modal information from the two data sets
Assessing public–private research collaboration: is it possible to compare university performance?
No full textIt is widely recognized that collaboration between the public and private research sectors should be stimulated and supported, as a means of favoring innovation and regional development. This work takes a bibliometric approach, based on co-authorship of scientific publications, to propose a model for comparative measurement of the performance of public research institutions in collaboration with the domestic industry collaboration with the private sector. The model relies on an identification and disambiguation algorithm developed by the authors to link each publication to its real authors. An example of application of the model is given, for the case of the academic system and private enterprises in Italy. The study demonstrates that for each scientific discipline and each national administrative region, it is possible to measure the performance of individual universities in both intra-regional and extra-regional collaboration, normalized with respect to advantages of location. Such results may be useful in informing regional policies and merit-based public funding of research organizations
- …
