Forward Calculation Research Articles

Distribution Center Location Model Based on Gauss-Kruger Projection and Gravity Method

The latitude and longitude coordinates are converted into Gaussian plane coordinates through the forward calculation of the Gauss-Kruger projection, and then the location is selected through the center of gravity method, and then the location result is converted into the latitude and longitude coordinates through the reverse calculation of the Gauss-Kruger projection. This method was then applied to the location of distribution centers among the four cities in Zhejiang Province, and reasonable results were obtained.

The role of thermomechanical processing routes on the grain boundary network of martensite in Ti–6Al–4V

The influence of thermomechanical processing (TMP) of a Ti–6Al–4V alloy on the transformation texture and intervariant boundary network were investigated by conventional EBSD mapping along with the five-parameter boundary analysis approach. The texture characteristics of Ti–6Al–4V subjected to deformation in the β regime followed by the β→α martensitic transformation were examined using visco-plastic self-consistent simulation and forward calculation of the transformation texture. Comparison of the simulated and experimental texture characterisitcs revealed that the transformed α′ texture was dominated by the variant selection associated with substructure development in the β parent phase and the occurrence of specific self-accommodating α′ variants in the microstructure, promoting the quadrilateral and/or V shape variant arrangement. This resulted in a progressive increase in the 63.26°/[10‾553‾] intervariant boundaries with the strain increment, at the expense of 60°/[112‾0]. Moreover, the grain boundary network for all conditions was dominated by the triple junctions (grain boundary network) terminating on 63.26°/[10‾553‾] and 60°/[112‾0] intervariants. It is shown that the elastic interactions among the variants during the martensitic transformation is the dominant parameter affecting the grain boundary network, despite the presence of dislocation based variant selection.

Computational Intelligence-based Evaluation of a 3-DOF Robotic-arm Forward Kinematics

Robotic manipulator- forward Kinematics involves the assurance of end-effector arrangements from connecting joint boundaries. The traditional mathematical calculation of controller forward -Kinematics is monotonous and tedious. Accordingly, it is important to execute a strategy that precisely performs forward energy while wiping out the disadvantages of the mathematical calculation technique. Versatile Neuro-Fuzzy Inference System (ANFIS) is a computational knowledge strategy that has been effectively executed for expectation purposes in assorted logical orders. This present examination's essential goal was to evaluate the productivity of ANFIS in foreseeing 3-levels of opportunity automated controller Cartesian directions from connecting joint boundaries. A speculative 3-level of opportunity automated controller has been considered in this investigation. Model preparing information has been obtained by mathematical forward kinematics calculation of the controller's end effector arrangements. Nine datasets have been utilized for model preparing, while five datasets have been utilized for model testing or approval. The ANFIS model's precision has been surveyed by figuring the Mean outright Percentage Error (MAPE) between the real and anticipated end-effector Cartesian directions. Because of Mean Absolute Percentage Error (MAPE), the created ANFIS model has forecast correctness’s of 63.35% and 80.07% in foreseeing x-directions and y-organizes, separately. Accordingly, ANFIS can be dependably executed as a commendable substitute for the customary arithmetical calculation method in anticipating controller Cartesian directions. It is suggested that the precision of other computational knowledge methods like Particle Swarm Optimization (PSO) and Support Vector Machines (SVM) be evaluated.

Sensitivity – Local index to control chaoticity or gradient globally –

Here, we introduce a fully local index named “sensitivity” for each neuron to control chaoticity or gradient globally in a neural network (NN). We also propose a learning method to adjust it named “sensitivity adjustment learning (SAL)”. The index is the gradient magnitude of its output with respect to its inputs. By adjusting its time average to 1.0 in each neuron, information transmission in the neuron changes to be moderate without shrinking or expanding for both forward and backward computations. That results in moderate information transmission through a layer of neurons when the weights and inputs are random. Therefore, SAL can control the chaoticity of the network dynamics in a recurrent NN (RNN). It can also solve the vanishing gradient problem in error backpropagation (BP) learning in a deep feedforward NN or an RNN. We demonstrate that when applying SAL to an RNN with small and random initial weights, log-sensitivity, which is the logarithm of RMS (root mean square) sensitivity over all the neurons, is equivalent to the maximum Lyapunov exponent until it reaches 0.0. We also show that SAL works with BP or BPTT (BP through time) to avoid the vanishing gradient problem in a 300-layer NN or an RNN that learns a problem with a lag of 300 steps between the first input and the output. Compared with manually fine-tuning the spectral radius of the weight matrix before learning, SAL’s continuous nonlinear learning nature prevents loss of sensitivities during learning, resulting in a significant improvement in learning performance.

Competitive Neural Network Circuit Based on Winner-Take-All Mechanism and Online Hebbian Learning Rule

In this article, we design a memristive competitive neural network circuit based on the winner-take-all (WTA) mechanism and the online Hebbian learning rule. Each synapse of the network contains two memristors whose terminals of signal inputs are opposite. However, only one memristor participates in the calculation each time, and that one is determined by the original input signal. The competitive neural network circuit includes two parts: forward calculation and weight update. In this article, the forward calculation part of the circuit is designed based on the WTA mechanism. The combination of the leaky-integrate-and-fire (LIF) model and pMOS realizes the lateral inhibition of neurons. The design of the weight updating part is based on Hebbian learning rules. In each cycle, only synaptic memristors connected to the winner output neuron in forward calculation can be adjusted. The voltage used for synaptic memristor adjustment comes from the membrane voltage of the winner output neuron. The whole neural network circuit does not need the participation of a central processing unit (CPU) or a field-programmable gate array (FPGA) and really realizes parallel calculation, the saving of area, power consumption, and a certain extent computing-in-memory. Based on the circuit designed in PSPICE, we simulated the classification of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$5\times3$ </tex-math></inline-formula> pixel pictures. The changing trend of weights in the training phase and the high recognition accuracy in the recognition phase prove that the network can learn and recognize different patterns. The competitive neural network can be applied to the neuromorphic system of visual pattern recognition.

A self-alignment method for gravitational apparent acceleration identification and accelerometer bias estimation based on repeated navigation solution.

Estimating inertial measurement unit error without any external reference information has always been a difficult problem. The method based on the reconstruction of the gravitational apparent motion in the inertial frame can estimate the accelerometer bias, but a long identification time is needed to get high alignment accuracy. In order to reduce the convergence time and improve the accuracy of estimation, a fast estimation method based on repetitive navigation under the excitation of swing motion is proposed. Two data processing methods, the forward-forward loop and the backward-forward loop, are studied and compared in this paper. In addition, the influence of data length on the experiment result was analyzed. The simulation and experimental results show that the forward-forward loop calculation cannot solve the contradiction between speed and accuracy, while the backward-forward loop calculation can effectively estimate the accelerometer bias and realize self-alignment, with a short alignment time and high alignment accuracy.

Research on Image Super-resolution Internet Transmission Technology Based on Virtual Optics

Image resolution is a set of performance parameters used to evaluate the richness of detailed information contained in an image, including temporal resolution, spatial resolution and color scale resolution, etc., which reflects the ability of the imaging system to reflect the detailed information of objects. In this paper, a super-resolution image transmission technology based on virtual optics is proposed. The original image is placed on the object plane of the virtual light path. Firstly, the light wave signal of the original image on the virtual diffraction plane is obtained by inverse diffraction calculation. Then, the light wave of the virtual diffraction plane is illuminated by spherical wave and then the forward diffraction calculation is performed. By changing the position of the observation plane, the original images with different magnification can be reconstructed. The simulation test results show that compared with the general interpolation amplification method, the amplified image has a good visual perception effect, especially in the significant area. The results obtained by the algorithm in this paper are better than the interpolation model and convolutional neural network algorithm as a whole; The above test results objectively show that the super-resolution effect of this algorithm is better.

Track Prediction for HF Radar Vessels Submerged in Strong Clutter Based on MSCNN Fusion with GRU-AM and AR Model

High-frequency (HF) surface-wave radar has a wide range of applications in marine monitoring due to its long-distance, wide-area, and all-weather detection ability. However, the accurate detection of HF radar vessels is severely restricted by strong clutter and interference, causing the echo of vessels completely submerged by clutter. As a result, the target cannot be detected and tracked for a period of time under the influence of strong clutter, which causes broken trajectories. To solve this problem, we propose an HF radar-vessel trajectory-prediction method based on a multi-scale convolutional neural network (MSCNN) that combines a gated recurrent unit and attention mechanism (GRU-AM) and a fusion with an autoregressive (AR) model. The vessel’s latitude and longitude information obtained by the HF radar is sent into the convolutional neural network (CNN) with different window lengths in parallel, and feature fusion is performed on the extracted multi-scale features. The deep GRU model is built to learn the time series with the GRU structure to preserve historical information. Different weights are given to the features using the temporal attention mechanism (AM), which helps the network learn the key information. The linear information on latitude and longitude at the current timestep is forecast by combining the AR model with the trajectory output from the AM to achieve a combination of linear and nonlinear prediction models. To make full use of the HF radar tracking information, the broken trajectory prediction is carried out by forward and backward computation using data from before and after the fracture, respectively. Weights are then assigned to the two predicted results by the entropy-value method to obtain the final ship trajectory by weighted summation. Field experiments show that the proposed method can accurately forecast the trajectories of vessels concealed in clutter. In comparison with other mainstream methods, the new method performs better in estimation accuracy for HF radar vessels concealed in clutter.

Comparison of Single-Site Ionospheric Delays Derived from Geometry-Free and Geometry-Based Approaches with an Analysis of Orbit and Clock Error Effects

The reparameterization of the geometry-free and geometry-based approaches to derive single-site ionospheric delays using Global Navigation Satellite System (GNSS) measurements is described. Kalman filtering is used to compute the geometry-free and geometry-based ionospheric delays in a forward computation procedure, aiming for a real-time application case. The numerical similarity and differences between the geometry-free and geometry-based ionospheric delays are assessed in terms of both formal and experimental errors (precision). The differences between geometry-free and geometry-based ionospheric delays are derived using two types of precise orbit and clock products. The effects of the precise orbit and clock residual errors are analyzed. The correlation coefficients between the L1 and L2 wide-lane ambiguities with the ionospheric delay are derived and analyzed. It is discovered that the geometry-based ionospheric delay is negatively correlated with geometry-based wide-lane ambiguities, while the geometry-free ionospheric delay and wide-lane ambiguities are much less correlated. A simulation analysis indicates that the impacts on geometry-based ionospheric delay estimates are partly coincided with the actual time-variant errors of the used orbit and clock in the line-of-sight direction.

FRoG dose computation meets Monte Carlo accuracy for proton therapy dose calculation in lung.

To benchmark and evaluate the clinical viability of novel analytical GPU-accelerated and CPU-based Monte Carlo (MC) dose-engines for spot-scanning intensity-modulated-proton-therapy (IMPT) towards the improvement of lung cancer treatment. Nine patient cases were collected from the CNAO clinical experience and The Cancer Imaging Archive-4D-Lung-Database for in-silico study. All plans were optimized with 2 orthogonal beams in RayStation (RS) v.8. Forward calculations were performed with FRoG, an independent dose calculation system using a fast robust approach to the pencil beam algorithm (PBA), RS-MC (CPU for v.8) and general-purpose MC (gp-MC). Dosimetric benchmarks were acquired via irradiation of a lung-like phantom and ionization chambers for both a single-field-uniform-dose (SFUD) and IMPT plans. Dose-volume-histograms, dose-difference and γ-analyses were conducted. With respect to reference gp-MC, the average dose to the GTV was 1.8% and 2.3% larger for FRoG and the RS-MC treatment planning system (TPS). FRoG and RS-MC showed a local γ-passing rate of ~96% and ~93%. Phantom measurements confirmed FRoG's high accuracywith a deviation<0.1%. Dose calculation performance using the GPU-accelerated analytical PBA, MC-TPS and gp-MC code were well within clinical tolerances. FRoG predictions were in good agreement with both the full gp-MC and experimental data for proton beams optimized for thoracic dose calculations. GPU-accelerated dose-engines like FRoG may alleviate current issues related to deficiencies in current commercial analytical proton beam models. The novel approach to the PBA implemented in FRoG is suitable for either clinical TPS or as an auxiliary dose-engine to support clinical activity for lung patients.

Line-of-Sight Winds and Doppler Effect Smearing in ACE-FTS Solar Occultation Measurements

Line-of-sight wind profiles are derived from Doppler shifts in infrared solar occultation measurements from the Atmospheric Chemistry Experiment Fourier transform spectrometers (ACE-FTS), the primary instrument on SCISAT, a satellite-based mission for monitoring the Earth’s atmosphere. Comparisons suggest a possible eastward bias from 20 m/s to 30 m/s in ACE-FTS results above 80 km relative to some datasets but no persistent bias relative to other datasets. For instruments operating in a limb geometry, looking through a wide range of altitudes, smearing of the Doppler effect along the line of sight can impact the measured signal, particularly for saturated absorption lines. Implications of Doppler effect smearing are investigated for forward model calculations and volume mixing ratio retrievals. Effects are generally small enough to be safely ignored, except for molecules having a large overhang in their volume mixing ratio profile, such as carbon monoxide.

Forward flow and creative thought: Assessing associative cognition and its role in divergent thinking

Creative thinking is thought to be supported by both spontaneous associative and controlled executive processes. Recently, a new measure of associative cognition has been developed—forward flow—which uses computational semantic models (e.g., latent semantic analysis; LSA) to capture “how far” people travel in semantic space during a chained free association task. The present research aims to extend the psychometrics of forward flow by 1) leveraging multiple computational semantic models for forward flow computation (reliability) and 2) testing how this metric of associative ability relates to divergent creative thinking (validity). In addition, using structural equation modeling, we test dual-process theories of creative cognition by examining the relative contribution of associative and executive abilities to divergent thinking. Study 1 (n = 151) finds moderately improved reliability of forward flow using the new multi-model approach (compared to LSA only), as well as positive effects of both forward flow (β = .48) and general intelligence (β = .36) on divergent thinking (human creativity ratings) in the same structural regression model. This pattern of results was replicated in Study 2 (n = 150), which showed large effects of forward flow (β = .42) and general intelligence (β = .46) on divergent thinking. The results expand the psychometrics of forward flow and provide new evidence for dual process models of creative cognition.

DEVELOPMENT OF INDUCED SEISMICITY IN MINING TASHTAGOL IRON ORE DEPOSIT: RETROSPECTIVE ANALYSIS AND FORECAST

Using the example of the Tashtagol iron ore deposit, an approach has been implemented that makes it possible to reveal the space-time relationships between the parameters of random and deterministic processes of various physical nature, which occur in the rock mass during large-scale mining operations. The approach consists in cross correlation analysis of long time series “number and energy of weak and moderate dynamic events” and “variation of stress field parameters” in different areas of geomechanical space of the object under study. According to the data of automated microseismic control system of the mine (from 1989 to 2019) and calculation results based on the developed 3D model of the deposit, the values of correlation coefficients r between number N of seismic events with an energy of up to 3 kJ and increments of the maximum tangential T and average S stresses in different blocks were determined. The regression dependences N(T) and N(S) constructed for statistically significant r were verified by comparing the predicted and real values of N in 2020. The obtained dependences can be used for geomechanical justification of reserves mining using forward calculations of the stress state and, on this basis, for assessing seismic activity level using the found correlation dependences.

Lane Line Detection Based on Object Feature Distillation

In order to meet the real-time requirements of the autonomous driving system, the existing method directly up-samples the encoder’s output feature map to pixel-wise prediction, thus neglecting the importance of the decoder for the prediction of detail features. In order to solve this problem, this paper proposes a general lane detection framework based on object feature distillation. Firstly, a decoder with strong feature prediction ability is added to the network using direct up-sampling method. Then, in the network training stage, the prediction results generated by the decoder are regarded as soft targets through knowledge distillation technology, so that the directly up-samples branch can learn more detailed lane information and have a strong feature prediction ability for the decoder. Finally, in the stage of network inference, we only need to use the direct up-sampling branch instead of the forward calculation of the decoder, so compared with the existing model, it can improve the lane detection performance without additional cost. In order to verify the effectiveness of this framework, it is applied to many mainstream lane segmentation methods such as SCNN, DeepLabv1, ResNet, etc. Experimental results show that, under the condition of no additional complexity, the proposed method can obtain higher F1Measure on CuLane dataset.

ESTIMATION OF SUBSURFACE STRUCTURE BASED ON MICROTREMOR AND SEISMIC OBSERVATIONS IN AREA DAMAGED BY 2018 HOKKAIDO EASTERN IBURI EARTHQUAKE, HOKKAIDO, JAPAN

To investigate the cause of serious damage during the 2018 Hokkaido Eastern Iburi Earthquake, we observed microtremors and aftershocks at a landslide area and around strong ground motion observation stations. Subsurface velocity structures were determined using a heuristic approach based on a forward calculation using the phase velocities and H/V spectra. As a result, S-wave velocity structures were estimated, and the predominant frequencies of the H/V spectra were obtained. The predominant frequencies of the H/V spectra of the microtremor and seismic waves are 0.8–5 Hz. The S-wave velocity of the uppermost layer is 75–130 m/s, and its thickness is approximately 10 m. In particular, it has been suggested that the uppermost layer of volcanic ash caused the landslide in the town of Atsuma. The predominant frequency corresponds to the thickness of the alluvial layers with S-wave velocities of 75–300 m/s. It was also found that the predominant direction of the horizontal earthquake ground motion of the landslide was affected by the topography of the mountain.

Realizing transmitted metasurface cloak by a tandem neural network

Being invisible at will has been a long-standing dream for centuries, epitomized by numerous legends; humans have never stopped their exploration steps to realize this dream. Recent years have witnessed a breakthrough in this search due to the advent of transformation optics, metamaterials, and metasurfaces. However, the previous metasurface cloaks typically work in a reflection manner that relies on a high-reflection background, thus limiting the applications. Here, we propose an easy yet viable approach to realize the transmitted metasurface cloak, just composed of two planar metasurfaces to hide an object inside, such as a cat. To tackle the hard-to-converge issue caused by the nonuniqueness phenomenon, we deploy a tandem neural network (T-NN) to efficiently streamline the inverse design. Once pretrained, the T-NN can work for a customer-desired electromagnetic response in one single forward computation, saving a great amount of time. Our work opens a new avenue to realize a transparent invisibility cloak, and the tandem-NN can also inspire the inverse design of other metamaterials and photonics.

An approach to determine radiative properties of solid struts of open-cell foams by pore scale identification from macro scale measurements

Open-cell foams consisting of interconnected struts are widely used in thermal applications such as volumetric solar receivers. The thermal performance of foams strongly relies on their radiative behaviors which intrinsically depends on the radiative properties of struts. This study proposed a feasible approach to determine the radiative properties of solid struts by a pore scale identification from a macro scale measurement. First, the normal-hemispherical transmittance and reflectance spectra of two nickel foams were measured by an experimental system. Second, a pore scale forward calculation was conducted by a combination of Monte Carlo Ray-tracing method and real foam structures obtained from micro-computed tomography technique. Last, an Adaptive Particle Swarm Optimization algorithm was adopted to retrieve the spectral reflectivity and specularity of nickel struts from the measured spectra. The results show that it is unreasonable to assume the reflection behavior of strut surface simply as either specular or diffused, and also, it is unreliable to assign the reflectivity of strut surface directly by Fresnel's law. Within the investigated wavebands 0.4–2.2 μm, the average reflectivity of nickel struts is 12% lower than that of a nickel mirror. The proportion of specular reflection at nickel strut surface is between 37%–52%, almost increasing with wavelengths.

Neutron noise simulator based on the boundary element method (BEM)

The purpose of the present study is to develop the neutron diffusion solver and neutron noise simulator based on the Boundary Element Method (BEM). The 2-D, 2-G neutron/adjoint diffusion equation and corresponding neutron/adjoint noise equation were solved using the mentioned method. The developed neutron static and noise simulator based on the finite element method gives accurate results when the more number of the elements is used. The motivation of the present research is to use the boundary element method to reduce the computational cost. The boundary element method attempts to use the givenboundary conditionsto fit boundary values into the integral equation, rather than values throughout the space defined by a partial differential equation.The results of static calculation were then benchmarked against the valid data for BIBLIS-2D and IAEA-2D benchmark problems. The neutron noise calculation due to the absorber of variable strength neutron noise source was performed using BEM and the Green's function technique. The adjoint of noise equation was also solved using the same method applied to the forward calculation. Comparison of the calculated neutron noise at zero frequency with the results of static calculation, utilizing the result of adjoint noise calculation and qualitative comparison with similar reported results are three different procedures to validate the neutron noise calculation.The analysis of the results obtained from the BEM shows that the boundary element method alone, or coupled with the finite element method, is ideally suited for the static and neutron noise calculation from both the accuracy and efficiency viewpoints.

Two-dimensional meshless modelling and TE-mode inversion of magnetotelluric data

SUMMARY We present a new 2-D forward modelling and inversion scheme to interpret magnetotelluric/radio-magnetotelluric data by utilizing a novel meshless forward operator. We use this discretization technique within an inverse scheme to recover conductivity structures from given magnetotelluric data. To approximate solutions of the partial differential equations that describe the magnetotelluric experiment, we discretize the subsurface only in terms of nodes. These node sets, which are simple to generate, are used to derive the differential operators’ approximations in a generalized meshless framework. First, we study and compare forward modelling calculations to an analytical and known solution from the literature. Several example calculations are given, which validate the proposed meshless forward operator. We then formulate our inverse scheme for TE-mode data, which uses only subsets of the nodal subsurface parametrization to generate conductivity structures from this given data. The inverse scheme consists of a Gauss–Newton algorithm combined with the generalized meshless framework. To validate the algorithm, we present inversion results from synthetic and field data. We compare our results to conductivity models calculated by established, well-known inversion schemes and literature results. We report that our algorithm can accurately model magnetotelluric responses and recover meaningful conductivity models, explaining given magnetotelluric data.

Reverse Calculation-Based Low Memory Turbo Decoder for Power Constrained Applications

Turbo codes are a family of near Shannon limit error correction coding schemes that usually are adopted for wireless data transmission. To reduce the power dissipation of a long-term evolution (LTE) advanced turbo decoder, in this paper, we propose a reverse calculation based low memory turbo decoder architecture by partitioning the trellis diagram and simplifying the max* operator. The designed forward state metrics calculation architecture is merged with two classical decoding schemes. Through field programmable gate array (FPGA) hardware implementation, the state metrics cache (SMC) capacity is reduced by 65%, the power dissipation of the reverse calculation architecture is significantly reduced for all tested clock frequencies, and the decoding performance is not affected as compared with classical decoding schemes. The proposed reverse calculation architecture is an effective technique to achieve better decoding performance for power-constrained applications.