Search results for "Approximation error"
showing 10 items of 62 documents
2014
A 3D imaging technique using a high speed binocular stereovision system was developed in combination with corresponding image processing algorithms for accurate determination of the parameters of particles leaving the spinning disks of centrifugal fertilizer spreaders. Validation of the stereo-matching algorithm using a virtual 3D stereovision simulator indicated an error of less than 2 pixels for 90% of the particles. The setup was validated using the cylindrical spread pattern of an experimental spreader. A 2D correlation coefficient of 90% and a Relative Error of 27% was found between the experimental results and the (simulated) spread pattern obtained with the developed setup. In combin…
A circuit model of a 5kW fuel cell
2013
In the last few years, industrial research has been focused on renewable energy sources. Owing to combined heat-power generation option, fuel cells have attracted much interest for a wide variety of research areas. Household appliances are still the most promising applications. Fuel cell modeling is still a critical topic to focus on, affecting each stage of power system design. This paper describes an empirical method to model a proton exchange membrane fuel cell. The modeling approach is applied to a 5kW Nuvera PowerFlow fuel cell. A circuit model is proposed for further applications. PSPICE implementation is described and the parameters tuning procedure is discussed. Simulation and exper…
Approximation properties of higher degree F-transforms based on B-splines
2015
The paper deals with the F-transform with polynomial components with respect to a generalized fuzzy partition given by B-splines. We investigate approximation properties of the inverse F-transform in this case and prove that using B-splines allows us to improve the quality of approximation of smooth functions.
Progressive Stochastic Binarization of Deep Networks
2019
A plethora of recent research has focused on improving the memory footprint and inference speed of deep networks by reducing the complexity of (i) numerical representations (for example, by deterministic or stochastic quantization) and (ii) arithmetic operations (for example, by binarization of weights). We propose a stochastic binarization scheme for deep networks that allows for efficient inference on hardware by restricting itself to additions of small integers and fixed shifts. Unlike previous approaches, the underlying randomized approximation is progressive, thus permitting an adaptive control of the accuracy of each operation at run-time. In a low-precision setting, we match the accu…
Fast Graph Filters for Decentralized Subspace Projection
2020
A number of inference problems with sensor networks involve projecting a measured signal onto a given subspace. In existing decentralized approaches, sensors communicate with their local neighbors to obtain a sequence of iterates that asymptotically converges to the desired projection. In contrast, the present paper develops methods that produce these projections in a finite and approximately minimal number of iterations. Building upon tools from graph signal processing, the problem is cast as the design of a graph filter which, in turn, is reduced to the design of a suitable graph shift operator. Exploiting the eigenstructure of the projection and shift matrices leads to an objective whose…
A simple method to directly retrieve reference evapotranspiration from geostationary satellite images
2013
Abstract Application of FAO-56 methodology for the assessment of reference evapotranspiration, ET 0 , is challenging in areas of the world with sparse meteorological network stations. For this reason alternative procedures using remotely observed data have been proposed in the literature. In this work, a simplified version of the Makkink approach [J. Inst. Wat. Eng. 11: 277–288, 1957] was tested in a typical Mediterranean environment (Sicily, Italy). The implemented Makkink approach (MAK) uses remotely estimated solar radiation derived from Meteosat Second Generation (MSG) satellite data and in situ observations of air temperature to assess ET 0 at daily time scale. Alternatively, taking ad…
Error and Uncertainty Analysis of the Residual Stresses Computed by Using the Hole Drilling Method
2010
: The hole-drilling method is one of the most used techniques for the experimental analysis of the residual stresses in mechanical components. For both through-thickness uniform and non-uniform residual stress distributions, its application is standardised by the ASTM E837-08. In accordance with the ASTM limitations, the analysis of uniform residual stresses, to which the present work deals with, leads in general to results with a maximum bias of about 10%. Unfortunately, in general the user does not have appropriate procedures to estimate the actual stress error; consequently, if one or more of the experimental influence parameters fall out of the corresponding standard limitations, the c…
Testing the BEST procedure to estimate the soil water retention curve
2012
The BEST (Beerkan Estimation of Soil Transfer parameters) procedure is attractive for simple soil hydraulic characterization but there is the need to test the reliability of the predictions. In this investigation, the BEST procedure to predict water retention of 199 Sicilian soils was evaluated. The BEST water retention model performed well (relative error, Er≤0.05) for approximately 80% of the soil samples. Low errors were obtained in soils with a high clay, cl, content (≥44%), whereas both high and low Er values were obtained in soils with a lower cl content. The BEST particle size distribution (PSD) model was accurate for 50% of the samples and the fitting accuracy increased with cl, wit…
Enabling XCSF to cope with dynamic environments via an adaptive error threshold
2020
The learning classifier system XCSF is a variant of XCS employed for function approximation. Although XCSF is a promising candidate for deployment in autonomous systems, its parameter dependability imposes a significant hurdle, as a-priori parameter optimization is not feasible for complex and changing environmental conditions. One of the most important parameters is the error threshold, which can be interpreted as a target bound on the approximation error and has to be set according to the approximated function. To enable XCSF to reliably approximate functions that change during runtime, we propose the use of an error threshold, which is adapted at run-time based on the currently achieved …
On Equivalent Random Traffic method extension
2011
The key result of the paper is the Equivalent Random Traffic (ERT) method extension for estimation of the throughput for schemes with traffic splitting. The excellent accuracy (relative error is less than 1%) is shown in numerical example. A numerical algorithm is given — how to estimate the throughput for schemes at traffic splitting and merging. The paper also contains new Erlang-B formula algorithm for non-integer number of channels based on parabolic approximation.