Abstract
Utility companies rely on solar radiation forecasting models to control the supply and demand of energy as well as the operability of the grid. They use these predictive models to schedule power plan operations, negotiate prices in the electricity market and improve the performance of solar technologies in general. This paper proposes a novel method for global horizontal irradiance forecasting. The method is based on an ensemble approach, in which individual competing models are arbitrated by a metalearning layer. The goal of arbitrating individual forecasters is to dynamically combine them according to their aptitude in the input data. We validate our proposed model for solar radiation forecasting using data collected by a real-world provider. The results from empirical experiments show that the proposed method is competitive with other methods, including current state-of-the-art methods used for time series forecasting tasks.

Abstract
We investigate spectroscopic imaging ellipsometry for monitoring biomolecules at surfaces of nanoparticles. For the modeling of polarimetric light scattering off surface-adsorbed core-shell nanoparticles, we employ an extension of the exact solution for the scattering by particles near a substrate presented by Bobbert and Vlieger, which offers insight beyond that of the Maxwell-Garnett effective medium approximation. Varying thickness and refractive index of a model bio-organic shell results in systematic and characteristic changes in spectroscopic parameters ? and ?. The salient features and trends in modeled spectra are in qualitative agreement with experimental data for antibody immobilization and fibronectin biorecognition at surfaces of gold nanoparticles on a silicon substrate, but achieving a full quantitative agreement will require including additional effects, such as nanoparticle-substrate interactions, into the model. © 2017 Optical Society of America.

Abstract

Abstract

Abstract
This study presents and applies a quantitative metric, based on entire time series of measured surface electromyography (sEMG) from selected lower limb muscles to validate multibody dynamics (MBD) estimated action of the same subject muscles during modified gait, stiff knee gait (SKG) and slow running (SR) in relation to normal gait (NG). MBD is being increasingly applied for estimation of internal actions according to difficulty of its direct measurements under natural conditions of movement and the importance of this estimation for prevention, diagnosis and treatment planning of specific subject skeletal and neuromuscular diseases. Inverse kinematics and inverse dynamics from position and force data have been used to estimate internal joint force moments, with muscle grouping and optimization techniques applied along with musculoskeletal model for estimation of muscle action. Nevertheless kinematic and kinetic input data of human movement must be accurate and employed model for simulation must be personalized to subject, task and moment of application. Also the results provided by the simulation with the musculoskeletal model must be compared with measured results for validation. Comparative analysis of kinematic and kinetic input data of human lower limbs is performed during modified gait modes and a personalized musculoskeletal model employed for MBD estimation of muscle actions and compare estimated muscle actions with measured sEMG of selected muscles on different gait modes, SKG and SR in relation to those registered at NG. The results from quantitative metrics followed qualitative agreement from visual inspection with better agreement between processed sEMG and MBD muscle estimated activity on phase metric than at magnitude, and combined metric presenting overall better agreement at NG and SKG than at SR, pointing to higher ability of the model to predict muscle force patterns in agreement with measured sEMG activity at NG and SKG than at SR and the need to improve model predictions for SR. Applied technique presents as a reproducible quantitative metric based on entire time series, both magnitude and phase, overcoming qualitative and subjective comparing by the observer, reducing time consuming and allowing increase at the number of automatic validation of MBD muscle action estimation.

Abstract
Public transport networks were, in the past, mainly designed to maximize the efficiency of commuting trips. However, with such perspective there are considerable risks to marginalize some specific population groups (e.g. disabled, elderly, children, pregnant, people in poverty). For enhancing social inclusion and improving the accessibility of more vulnerable citizens, such networks are often redesigned and adjusted. Nevertheless, even with such adjustments, it is sometimes difficult to provide efficient services that fully address the real needs and capabilities of travelers, partially because of the failure in following the fast technological and demanding changes of modern societies. Taking in mind these challenges, we have developed a conceptual model to support knowledge sharing and decision-making in urban mobility, and to improve the way travel information is addressed. The multi-user integrated platform proposed in this work is supported by the idea that information from different channels must be centralized, organized, managed and properly distributed. This idea is grounded in two main principles: (i) past and real-time information from a wide range of sources is combined for knowledge extraction, and such knowledge is going to be used not only to allow travelers to better plan their trips, but also to help transport providers to develop services adapted to the needs and preferences of their customers; and (ii) information is provided in a personalized way taking into account socio-economical differences between groups of travelers. (C) 2017 The Authors. Published by Elsevier B.V.