Higher RMS Error Research Articles

Determination of the thickness and optical properties by reflectance method

In this study, we propose a method based on the reflection spectrum envelope to calculate the thickness and optical constants of a semi-transparent thin film deposited on a glass substrate. To validate this method, we compare these optical constants with those obtained from the transmission spectrum in two different thickness cases: one involving a film with a thickness of 700 nm and the other with a thin film of 80 nm. Comparative results of the methods, including well-established approaches such as Swanepoel, Chambouleyron, Sellmeier, Lorentz and Forouhi–Bloomer reveal notable accuracy with RMS errors of 0.21, 0.22, 0.31, 0.36 and 0.38% respectively. In contrast, the “Cauchy” method shows lower precision with a higher RMS error at 0.71%. The method proposed in this study, applied to a film with a thickness of 700 nm, demonstrates the lowest RMS error at 0.17%.

Toward an affordable and user-friendly visual motion capture system.

The present study aims at designing and evaluating a low-cost, simple and portable system for arm joint angle estimation during grasping-like motions. The system is based on a single RGB-D camera and three customized markers. The automatically detected and tracked marker positions were used as inputs to an offline inverse kinematic process based on bio-mechanical constraints to reduce noise effect and handle marker occlusion. The method was validated on 4 subjects with different motions. The joint angles were estimated both with the proposed low-cost system and, a stereophotogrammetric system. Comparative analysis shows good accuracy with high correlation coefficient (r= 0.92) and low average RMS error (3.8 deg).

Model Stacking (MOST) technique applied in cross-hole ERT field data for the detection of Thessaloniki ancient walls' depth

Abstract A cross-hole Electrical Resistivity Tomography (ERT) study was undertaken near the center of Thessaloniki in order to detect the depth of the existing city walls in the planned route of the new city underground train. This cross-hole setup was used for a study of measurements with various electrode arrays in real urban field conditions to evaluate the resolution of the models which is produced by each array and the reliability of the models which is produced by the newly published “MOST” technique. The pole–tripole array (C2–C1P1P2) produces high resolution models, even when only borehole electrodes are used. The bipole–bipole C1C2–P1P2 array, when used for cross-hole measurements only, produces higher resolution models compared to the C1P1–C2P2 array, even with a lower signal-to-noise ratio, which can result in extremely high RMS error, when noise, systematic or not, must be faced. The models of both arrays are greatly improved by the use of surface electrodes. The pole–bipole array (C1–P1P2) is proved to be less accurate in imaging and quite unstable to the noisy urban environment and to systematic errors. Furthermore, the Model Stacking (MOST) interpretation technique leads to better results with models of greater resolution and fewer artifacts compared even with the combined data inversion. Finally, the ERT cross-hole analysis has been reliable in detecting the city walls.

Predicting the outcome of respiratory motion prediction

Prediction of respiratory motion traces has become an important research topic. Especially for motion compensated radiotherapy, compensation of the latencies arising from mechanical constraints and signal processing is necessary. In recent years, many algorithms have been developed and evaluated. It is, however, still unclear how well a specific patient will be suited to motion prediction before the treatment actually starts. In this work, we have analyzed 304 respiratory motion traces with an average duration of 71 min. A total of 21 features characterizing these signals (12 from the frequency domain and 9 from the time domain) have been determined for each motion trace. The correlation between these features and the overall prediction quality for three different algorithms (based on wavelet-based multiscale autoregression, support vector regression, and linear expansion of the prediction error) has been analyzed and six dominant features have been identified (three each from the time and frequency domains). Additionally, the optimized results of the multistep-linear method (MULIN) prediction algorithm on the first 300 s of motion data have been used as a seventh, independent feature. Assessing the prediction algorithms' quality was done by calculating the relative root mean squared (RMS(rel)) error, i.e., the ratio between the RMS error of the prediction output and the RMS error of the delayed signal (the RMS error obtained when doing no prediction). Then, for each algorithm, the signals themselves were grouped into four classes according to the quality of prediction: relative RMS less than 0.8 (C1), between 0.8 and 0.9 (C2), between 0.9 and 1.0 (C3), and over 1.0 (C4). The goal of this work is to identify, prior to treatment, those patients whose respiratory behavior indicates probable (RMS(rel) ≥ 0.9) or certain (RMS(rel) ≥ 1.0) failure of respiratory motion prediction. Consequently, all signals from C4 must be identified and rejected and no signals from C1 may be falsely rejected. The restriction on C2 and C3 is slightly weaker: C2 are those signals that should be kept and C3 are those signals that should be rejected. Rejecting all signals from C4 and C3, keeping as many signals from C1 and as few from C2 as possible, has been achieved for the wLMS algorithm when using six feature pairs and the result of prediction on the short signal. Here, the false rejectance rate for C1 was less than 13% and the false acceptance rate for C2 was 15%. For the SVRpred and MULIN algorithms, the results are somewhat worse: in both cases, signals from C3 were falsely accepted (25.0% and 14.3%, respectively) but all signals from C4 were rejected. The false rejectance rate for C1 was 11.4% (MULIN) and 26.3% (SVRpred). In general, it has been shown that pretreatment classification of the quality of respiratory motion prediction is possible and that signals with high relative RMS error can be identified with great reliability. This is especially true for the wLMS algorithm, which has also been identified as the most precise and robust of the presented methods.

Prediction of Raw Milk Microbial Quality Using Data Mining Techniques

Data mining techniques were applied to predict raw milk quality in terms of methylene blue reduction time (MBRT) from the independent parameters of raw milk inspection parameters such as travel time, temperature of milk, solid-not-fat, %fat, acidity and specific gravity. Predictive models were developed and the performance of 3 data mining algorithms namely; Multiple Linear Regression (MLR), Artificial Neural Network (ANN) and K-Nearest neighbor (KNN), was measured in terms of average error and Root Mean Square Error (RMSE). MLR showed high and inconsistent RMS error in 3 randomly picked data partitions whereas KNN and ANN were able to predict the MBRT values from the physico-chemical quality parameters, KNN was the preferred algorithm (K=7, RMSE of 1.7). The models were applied to a new set of data (n=78) without showing them the output parameter (MBRT). The predicted values of MBRT were plotted against the actual observed values to classify milk into 4 quality grades.

Simultaneous fitting of a potential-energy surface and its corresponding force fields using feedforward neural networks

An improved neural network (NN) approach is presented for the simultaneous development of accurate potential-energy hypersurfaces and corresponding force fields that can be utilized to conduct ab initio molecular dynamics and Monte Carlo studies on gas-phase chemical reactions. The method is termed as combined function derivative approximation (CFDA). The novelty of the CFDA method lies in the fact that although the NN has only a single output neuron that represents potential energy, the network is trained in such a way that the derivatives of the NN output match the gradient of the potential-energy hypersurface. Accurate force fields can therefore be computed simply by differentiating the network. Both the computed energies and the gradients are then accurately interpolated using the NN. This approach is superior to having the gradients appear in the output layer of the NN because it greatly simplifies the required architecture of the network. The CFDA permits weighting of function fitting relative to gradient fitting. In every test that we have run on six different systems, CFDA training (without a validation set) has produced smaller out-of-sample testing error than early stopping (with a validation set) or Bayesian regularization (without a validation set). This indicates that CFDA training does a better job of preventing overfitting than the standard methods currently in use. The training data can be obtained using an empirical potential surface or any ab initio method. The accuracy and interpolation power of the method have been tested for the reaction dynamics of H+HBr using an analytical potential. The results show that the present NN training technique produces more accurate fits to both the potential-energy surface as well as the corresponding force fields than the previous methods. The fitting and interpolation accuracy is so high (rms error=1.2 cm(-1)) that trajectories computed on the NN potential exhibit point-by-point agreement with corresponding trajectories on the analytic surface.

Multitemporal Analysis of TRMM-Based Satellite Precipitation Products for Land Data Assimilation Applications

Abstract In this study, the recent work of Gottschalck et al. and Ebert et al. is extended by assessing the suitability of two Tropical Rainfall Measuring Mission (TRMM)-based precipitation products for hydrological land data assimilation applications. The two products are NASA’s gauge-corrected TRMM 3B42 Version 6 (3B42), and the satellite-only NOAA Climate Prediction Center (CPC) morphing technique (CMORPH). The two products were evaluated against ground-based rain gauge–only and gauge-corrected Doppler radar measurements. The analyses were performed at multiple time scales, ranging from annual to diurnal, for the period March 2003 through February 2006. The analyses show that at annual or seasonal time scales, TRMM 3B42 has much lower biases and RMS errors than CMORPH. CMORPH shows season-dependent biases, with overestimation in summer and underestimation in winter. This leads to 50% higher RMS errors in CMORPH’s area-averaged daily precipitation than TRMM 3B42. At shorter time scales (5 days or less), CMORPH has slightly less uncertainty, and about 10%–20% higher probability of detection of rain events than TRMM 3B42. In addition, the satellite estimates detect more high-intensity events, causing a remarkable shift in precipitation spectrum. Summertime diurnal cycles in the United States are well captured by both products, although the 8-km CMORPH seems to capture more diurnal features than the 0.25° CMORPH or 3B42 products. CMORPH tends to overestimate the amplitude of the diurnal cycles, particularly in the central United States. Possible causes for the discrepancies between these products are discussed.

Global mineral distributions on Mars

Determining the mineralogy of Mars is an essential part of revealing the conditions of the surface and subsurface. A deconvolution method was used to remove atmospheric components and determine surface mineralogy from Thermal Emission Spectrometer data at 1 pixel per degree (ppd). Minerals are grouped into categories on the basis of compositional and spectral similarity, and global concentration maps are produced. All binned pixels are fit well with RMS errors of ≤0.005 in emissivity. Higher RMS errors are attributed to short wavelength particle size effects on dust‐covered surfaces. Significant concentrations (>0.10) of plagioclase, high‐Ca pyroxene, sheet silicates/high‐Si glass, and hematite are detected and display distributions consistent with previous studies. Elevated concentrations of plagioclase and high‐Ca pyroxene are consistent with basaltic surfaces and are located in low‐albedo highlands regions north of ∼45°S. Significant concentrations of plagioclase and sheet silicates/high‐Si glass and low concentrations of high‐Ca pyroxenes are consistent with andesitic surfaces and are concentrated in both southern and northern high‐latitude, low‐albedo regions. Andesitic surfaces in the southern hemisphere have a lower spectral contrast than northern surfaces. An isolated surface located in Solis Planum is spectrally distinct but compositionally similar to other surfaces interpreted to be andesitic in composition. Concentrations of olivine below the detection limit correctly identify its presence in two of three locations. Potassium feldspar, low‐Ca pyroxene, basaltic glass, olivine, sulfate, carbonate, quartz, and amphibole are not detected with confidence at 1 ppd. The results presented here indicate a predominance of volcanic compositions within Martian dust‐free surfaces.

A magnetospheric specification model validation study: Geosynchronous electrons

The Rice University Magnetospheric Specification Model (MSM) is an operational space environment model of the inner and middle magnetosphere designed to specify charged particle fluxes up to 100 keV. Validation test data taken between January 1996 and June 1998 consist of electron fluxes measured by a charge control system (CCS) on a defense satellite communications system (DSCS) spacecraft. The CCS includes both electrostatic analyzers to measure the particle environment and surface potential monitors to track differential charging between various materials and vehicle ground. While typical RMS error analysis methods provide a sense of the models overall abilities, they do not specifically address physical situations critical to operations, i.e., how well does the model specify when a high differential charging state is probable. In this validation study, differential charging states observed by DSCS are used to determine several threshold fluxes for the associated 20–50 keV electrons and joint probability distributions are constructed to determine Hit, Miss, and False Alarm rates for the models. An MSM run covering the two and one-half year interval is performed using the minimum required input parameter set, consisting of only the magnetic activity index Kp, in order to statistically examine the model's seasonal and yearly performance. In addition, the relative merits of the input parameter, i.e., Kp, Dst, the equatorward boundary of diffuse aurora at midnight, cross-polar cap potential, solar wind density and velocity, and interplanetary magnetic field values, are evaluated as drivers of shorter model runs of 100 d each. In an effort to develop operational tools that can address spacecraft charging issues, we also identify temporal features in the model output that can be directly linked to input parameter variations and model boundary conditions. All model output is interpreted using the full three-dimensional, dipole tilt-dependent algorithms currently in operational use at the Air Force 55th Space Weather Squadron (55 SWXS). Results indicate that both diurnal and seasonal activity related variations in geosynchronous electrons are reproduced in a regular and consistent manner regardless of the input parameter used as drivers. The ability of the MSM to specify DSCS electrons in relation to thresholds indicative of spacecraft charging varies with the combination of input parameters used. The input parameter of greatest benefit to the MSM, after the required Kp index, is the polar cap potential drop as determined by DMSP spacecraft. Regarding the highest electron flux threshold, the model typically achieves high HIT rates paired with both high False Alarm rates and higher RMS error. Suggestions are made regarding the utilization of proxy values for the polar cap potential parameter and Kp-dependent model boundary conditions. The importance of generating accurate real-time proxy input data for operational use is stressed.

A comparison of Sun photometer derivations of total column water vapor and ozone to standard measures of same at the Southern Great Plains Atmospheric Radiation Measurement site

Total column water vapor measurements made over a 1‐year period by collocated instruments at the Southern Great Plains Atmospheric Radiation Measurement program site in northern Oklahoma are compared. Microwave radiometer measurements were validated in an earlier program with extensive comparisons to balloon sonde measurements. Sun photometry, using the multifilter rotating shadowband radiometer (MFRSR), was the second technique used to derive water vapor using the 940‐nm water band when no clouds were between the radiometer and the sun. Bruegge et al.'s implementation of Reagan et al.'s modified Langley technique was applied in this case. This approach is noteworthy in that it does not rely on a comparison to a standard for calibration. The comparison of 101 one‐half‐hour to two‐hour averages produced a root‐mean‐square difference of 0.18 cm compared to a mean value of 1.49 cm, or about 12%, with the MFRSR showing an overall bias of +0.03 cm relative to the microwave radiometer data. This is a somewhat higher rms error than other results using the modified Langley technique, but the comparison was made over the entire range of water vapor columns experienced in a year of observations. As a by‐product of the water vapor derivation procedure for the MFRSR, ozone column abundance was estimated. Seasonally averaged, these data compare favorably with Dobson measurements from Nashville, Tennessee, which is near the same latitude. A more definitive conclusion regarding the Sun photometer ozone derivation technique requires colocated ozone measurements using an accepted standard approach, however, the technique appears to provide a useful ozone estimate for correcting aerosol optical depth derivations.